Lineage expression over time

#Set working directory to appropriate folder for inputs and outputs on Google Drive

#Initialize

Load data

rm(list = ls())
library(dplyr)
library(Seurat)
library(ggplot2)
library(RColorBrewer)
library(ggpubr)
library(pheatmap)
library(viridis)
library(xlsx)

`%nin%` = Negate(`%in%`)

need to make Idents metadata object which says if the cells are included in the combined lins list, or if they were filtered

load('2022_01_14_analysis_scripts/2022_05_27_analysis/Assign_dominant_barcodes/all_data_final_lineages.RData')
load('2022_01_14_analysis_scripts/2022_05_27_analysis/Preprocess_GEX/second_timepoint_merged.RData')
load('2022_01_14_analysis_scripts/2022_05_27_analysis/Filtering_cDNA/resistant_lineage_lists.RData')

load('2022_01_14_analysis_scripts/2022_05_27_analysis/Assign_dominant_barcodes/cis_final_lineages.RData')
load('2022_01_14_analysis_scripts/2022_05_27_analysis/Assign_dominant_barcodes/cocl2_final_lineages.RData')
load('2022_01_14_analysis_scripts/2022_05_27_analysis/Assign_dominant_barcodes/dabtram_final_lineages.RData')
load('2022_01_14_analysis_scripts/2022_05_27_analysis/Assign_dominant_barcodes/dabtram_both_times_final_lineages.RData')

Look at similarity within lineages based on pearson correlation

Look at whether lineages cluster together in each individual condition - Starting with DabTram

#find lineages that are maintained at both dabtram timepoints
fivecell_cDNA$DabTramMaintained <- Reduce(intersect, list(fivecell_cDNA$DabTram, fivecell_cDNA$DabTramtoDabTram))

filtered_meta <- rep(0, length(names(all_data$Lineage)))

#specify which cells are in lineages that pass filtering for that condition
filtered_meta[which(all_data$OG_condition == "dabtram" & all_data$Lineage %in% combined_lins_list$DabTram)] <- 'Resistant to DabTram'
filtered_meta[which(all_data$OG_condition == "dabtramtodabtram" & all_data$Lineage %in% combined_lins_list$DabTramtoDabTram)] <- 'Resistant to DabTramtoDabTram'
filtered_meta[which(all_data$OG_condition == "dabtramtococl2" & all_data$Lineage %in% combined_lins_list$DabTramtoCoCl2)] <- 'Resistant to DabTramtoCoCl2'
filtered_meta[which(all_data$OG_condition == "dabtramtocis" & all_data$Lineage %in% combined_lins_list$DabTramtoCis)] <- 'Resistant to DabTramtoCis'
filtered_meta[which(all_data$OG_condition == "cocl2" & all_data$Lineage %in% combined_lins_list$CoCl2)] <- 'Resistant to CoCl2'
filtered_meta[which(all_data$OG_condition == "cocl2todabtram" & all_data$Lineage %in% combined_lins_list$CoCl2toDabTram)] <- 'Resistant to CoCl2toDabTram'
filtered_meta[which(all_data$OG_condition == "cocl2tococl2" & all_data$Lineage %in% combined_lins_list$CoCl2toCoCl2)] <- 'Resistant to CoCl2toCoCl2'
filtered_meta[which(all_data$OG_condition == "cocl2tocis" & all_data$Lineage %in% combined_lins_list$CoCl2toCis)] <- 'Resistant to CoCl2toCis'
filtered_meta[which(all_data$OG_condition == "cis" & all_data$Lineage %in% combined_lins_list$Cis)] <- 'Resistant to Cis'
filtered_meta[which(all_data$OG_condition == "cistodabtram" & all_data$Lineage %in% combined_lins_list$CistoDabTram)] <- 'Resistant to CistoDabTram'
filtered_meta[which(all_data$OG_condition == "cistococl2" & all_data$Lineage %in% combined_lins_list$CistoCoCl2)] <- 'Resistant to CistoCoCl2'
filtered_meta[which(all_data$OG_condition == "cistocis" & all_data$Lineage %in% combined_lins_list$CistoCis)] <- 'Resistant to CistoCis'

#specify which cells are in lineages of more than 5 cells
filtered_meta[which(all_data$OG_condition == "dabtram" & all_data$Lineage %in% fivecell_cDNA$DabTram)] <- 'Large Resistant to DabTram'
filtered_meta[which(all_data$OG_condition == "dabtramtodabtram" & all_data$Lineage %in% fivecell_cDNA$DabTramtoDabTram)] <- 'Large Resistant to DabTramtoDabTram'
filtered_meta[which(all_data$OG_condition == "dabtramtococl2" & all_data$Lineage %in% fivecell_cDNA$DabTramtoCoCl2)] <- 'Large Resistant to DabTramtoCoCl2'
filtered_meta[which(all_data$OG_condition == "dabtramtocis" & all_data$Lineage %in% fivecell_cDNA$DabTramtoCis)] <- 'Large Resistant to DabTramtoCis'
filtered_meta[which(all_data$OG_condition == "cocl2" & all_data$Lineage %in% fivecell_cDNA$CoCl2)] <- 'Large Resistant to CoCl2'
filtered_meta[which(all_data$OG_condition == "cocl2todabtram" & all_data$Lineage %in% fivecell_cDNA$CoCl2toDabTram)] <- 'Large Resistant to CoCl2toDabTram'
filtered_meta[which(all_data$OG_condition == "cocl2tococl2" & all_data$Lineage %in% fivecell_cDNA$CoCl2toCoCl2)] <- 'Large Resistant to CoCl2toCoCl2'
filtered_meta[which(all_data$OG_condition == "cocl2tocis" & all_data$Lineage %in% fivecell_cDNA$CoCl2toCis)] <- 'Large Resistant to CoCl2toCis'
filtered_meta[which(all_data$OG_condition == "cis" & all_data$Lineage %in% fivecell_cDNA$Cis)] <- 'Large Resistant to Cis'
filtered_meta[which(all_data$OG_condition == "cistodabtram" & all_data$Lineage %in% fivecell_cDNA$CistoDabTram)] <- 'Large Resistant to CistoDabTram'
filtered_meta[which(all_data$OG_condition == "cistococl2" & all_data$Lineage %in% fivecell_cDNA$CistoCoCl2)] <- 'Large Resistant to CistoCoCl2'
filtered_meta[which(all_data$OG_condition == "cistocis" & all_data$Lineage %in% fivecell_cDNA$CistoCis)] <- 'Large Resistant to CistoCis'

# filtered_meta[which(all_data$OG_condition == "dabtram" & all_data$Lineage %in% fivecell_cDNA$DabTramMaintained)] <- 'Maintained Resistant to DabTram'
# filtered_meta[which(all_data$OG_condition == "dabtramtodabtram" & all_data$Lineage %in% fivecell_cDNA$DabTramMaintained)] <- 'Maintained Resistant to DabTramtoDabTram'

#specify which cells are in lineages that did not pass filtering
filtered_meta[which(all_data$OG_condition == "dabtram" & all_data$Lineage %nin% combined_lins_list$DabTram & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "dabtramtodabtram" & all_data$Lineage %nin% combined_lins_list$DabTramtoDabTram & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "dabtramtococl2" & all_data$Lineage %nin% combined_lins_list$DabTramtoCoCl2 & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "dabtramtocis" & all_data$Lineage %nin% combined_lins_list$DabTramtoCis & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cocl2" & all_data$Lineage %nin% combined_lins_list$CoCl2 & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cocl2todabtram" & all_data$Lineage %nin% combined_lins_list$CoCl2toDabTram & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cocl2tococl2" & all_data$Lineage %nin% combined_lins_list$CoCl2toCoCl2 & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cocl2tocis" & all_data$Lineage %nin% combined_lins_list$CoCl2toCis & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cis" & all_data$Lineage %nin% combined_lins_list$Cis & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cistodabtram" & all_data$Lineage %nin% combined_lins_list$CistoDabTram & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cistococl2" & all_data$Lineage %nin% combined_lins_list$CistoCoCl2 & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 
filtered_meta[which(all_data$OG_condition == "cistocis" & all_data$Lineage %nin% combined_lins_list$CistoCis & all_data$Lineage %nin% c("No Barcode", "Still multiple"))] <- 'Filtered out' 

#specify which cells had zero or multiple barcodes
filtered_meta[which(all_data$Lineage %in% c("No Barcode", "Still multiple"))] <- 'No Barcode'

print(table(filtered_meta))

filtered_meta
                       Filtered out              Large Resistant to Cis 
                               3337                                1375 
        Large Resistant to CistoCis       Large Resistant to CistoCoCl2 
                                951                                2078 
    Large Resistant to CistoDabTram            Large Resistant to CoCl2 
                               1394                                1784 
      Large Resistant to CoCl2toCis     Large Resistant to CoCl2toCoCl2 
                               3010                               11578 
  Large Resistant to CoCl2toDabTram          Large Resistant to DabTram 
                                663                                 478 
    Large Resistant to DabTramtoCis   Large Resistant to DabTramtoCoCl2 
                               4234                                2840 
Large Resistant to DabTramtoDabTram                          No Barcode 
                               3176                               35314 
                   Resistant to Cis               Resistant to CistoCis 
                                331                                  67 
            Resistant to CistoCoCl2           Resistant to CistoDabTram 
                                135                                 113 
                 Resistant to CoCl2             Resistant to CoCl2toCis 
                                278                                 157 
          Resistant to CoCl2toCoCl2         Resistant to CoCl2toDabTram 
                                 55                                  93 
               Resistant to DabTram           Resistant to DabTramtoCis 
                                337                                 225 
        Resistant to DabTramtoCoCl2       Resistant to DabTramtoDabTram 
                                100                                 222 

Significance testing of the dabtram simulation

Look at whether lineages cluster together in each individual condition - dabtramtodabtram

# Find the mean of the average pearson correlation per lineage
mean_pearson_dabtram <- mean(unlist(lapply(dabtram_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_dabtram_sim <- sapply(1:length(dabtram_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(dabtram_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_dabtram <- (mean_pearson_dabtram-mean(means_pearson_dabtram_sim))/sd(means_pearson_dabtram_sim) # Z score comparing mean to simulations
pval_mean_pearson_dabtram <- pnorm(z_mean_pearson_dabtram, mean(means_pearson_dabtram_sim), sd(means_pearson_dabtram_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_dabtram <- weighted.mean(unlist(lapply(dabtram_lin_pearson_list, mean)),
unlist(lapply(dabtram_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_dabtram_sim <- sapply(1:length(dabtram_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(dabtram_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(dabtram_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_dabtram <- (weighted_mean_pearson_dabtram-mean(weighted_means_pearson_dabtram_sim))/sd(weighted_means_pearson_dabtram_sim) # Z score comparing mean to simulations
pval_wmean_pearson_dabtram <- pnorm(z_wmean_pearson_dabtram, mean(weighted_means_pearson_dabtram_sim), sd(weighted_means_pearson_dabtram_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_dabtram <- c()
for (i in 1:length(dabtram_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(dabtram_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_dabtram <- cbind(wilcox_pval_dabtram, wilcox.test(x = unlist(lapply(dabtram_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(dabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties

# Save outputs
save(dabtram, dabtram_lin_pearson_list, dabtram_lin_pearson_rand_list, z_mean_pearson_dabtram, pval_mean_pearson_dabtram, z_wmean_pearson_dabtram, pval_wmean_pearson_dabtram,  wilcox_pval_dabtram, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/dabtram_pearson_sim_results.RData')
rm(dabtram,dabtram_lin_pearson_list, dabtram_lin_pearson_rand_list, dabtram_input_data)

Significance testing of the dabtramtodabtram simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to dabtramtodabtram
dabtramtodabtram <- subset(all_data, idents = 'dabtramtodabtram') # Subset down to the dabtramtodabtram object
dabtramtodabtram <- NormalizeData(dabtramtodabtram)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtodabtram <- FindVariableFeatures(dabtramtodabtram, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtodabtram <- ScaleData(dabtramtodabtram)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

dabtramtodabtram <- RunPCA(dabtramtodabtram)

PC_ 1 
Positive:  AEBP1, SORBS2, PSAP, PMP22, CALD1, CRYAB, CTSK, CEBPD, MALAT1, NUPR1 
       PLP1, NEAT1, DEPP1, FAM89A, SOX4, GAS7, SPARC, MBP, MLANA, PHACTR1 
       FBXO32, GPM6B, TXNIP, SOX10, S100B, CBLB, PDZRN3, ID4, AKAP12, PMEL 
Negative:  MKI67, CENPF, TPX2, STMN1, RRM2, UBE2C, PCLAF, TOP2A, ANLN, UBE2S 
       BIRC5, H2AFZ, TK1, CKS1B, GTSE1, PRC1, NUSAP1, HIST1H4C, UBE2T, HIST1H1B 
       DTYMK, ASPM, TUBB, TMPO, TUBA1B, CEP55, HIST1H1D, FOSL1, ZWINT, RANBP1 
PC_ 2 
Positive:  VCAM1, COL14A1, IGFBP5, C1R, APOE, TXNRD1, COL1A1, COL6A1, CTSD, PDE5A 
       ADM, SLC7A11, IL6ST, SQSTM1, RND3, IRF1, HLA-B, LINC00968, DCN, LAMB1 
       PLAAT4, NQO1, LUM, BOC, CTSC, NFE2L2, CARD16, WNT5A, MGST1, LBP 
Negative:  CRYAB, RPL26, S100B, FABP5, MIF, MT2A, PLP1, S100A1, GAS7, PHACTR1 
       SH3BGRL3, MLANA, H2AFZ, SOX10, LIMA1, C3orf14, S100A10, CHPF, H2AFJ, C12orf75 
       TAGLN2, RNASEK, COL9A3, CSTB, GPM6B, ITGA6, NGFR, VGF, RPL22L1, UBE2C 
PC_ 3 
Positive:  TMSB4X, C12orf75, CAV1, ARL4C, CPA4, CEMIP, RPS27L, DKK1, S100A6, CTHRC1 
       SERPINE1, DSTN, S100A11, RPL26, SFRP1, AC092807.3, DAB2, TIMP3, CSTB, RPL22L1 
       LMO7, S100A16, TNIK, RNASEK, SMYD3, SRGN, SPOCK1, CCL2, RPL12, LINC01638 
Negative:  GAS7, CRYAB, PLP1, COL9A3, SOX10, AKAP12, PHACTR1, NGFR, PMP22, MLANA 
       S100B, TXNIP, SOX6, ITGA6, CTSD, PMEPA1, EPB41L3, FXYD3, GPM6B, PMP2 
       NFATC2, EPB41L2, TSC22D1, EMP1, AL139383.1, TRIM2, TFAP2A, S100A1, CITED1, MGP 
PC_ 4 
Positive:  MALAT1, HSP90B1, FN1, MT-ND1, MT-ATP6, SERPINE2, MT-ND2, MT-ND4, TIMP3, SPARC 
       MFGE8, NEAT1, CALR, THBS1, SPOCK1, LMO7, HSPA5, GLS, IGFBP7, LIMA1 
       ANXA2, HSPA8, EIF4A1, PRNP, SORBS2, MYH9, PHLDB2, COL5A2, PRSS23, MT-ND6 
Negative:  FTH1, FTL, MIF, TXN, RPL12, RPL26, MGST1, S100A6, CSTB, GYPC 
       ATOX1, RPL34, TOMM5, ATP5MC1, PTGR1, RPS27L, CENPX, RNASEK, ITGAE, H2AFZ 
       KYNU, SOST, CHCHD10, GAPDH, CYTL1, AKR1B10, AKR1C3, BOC, JPT1, MMD 
PC_ 5 
Positive:  SH3BGRL3, S100B, S100A10, GAS7, ATOX1, CENPX, COL9A3, GYPC, CLSPN, PHLDA2 
       HSPG2, IFI27, NGFR, ATP5MC1, CRYAB, E2F1, CALM2, PMP2, S100A11, CALD1 
       DTL, COTL1, HELLS, ATAD2, CCNE2, DKK1, S100A6, EPB41L3, SOX6, ITGA2 
Negative:  CTSK, CCNB1, PMEL, PSAP, ASPM, CENPE, HMMR, PLK1, MT-ATP6, AURKA 
       DLGAP5, FBXO32, CTSL, CENPF, LGALS3, NEK2, CCNB2, IGFN1, CDC20, APOE 
       ARL6IP1, KIF14, TPX2, CENPA, DEPDC1, MT-ND2, PTTG1, TSPAN10, CKAP2, SLC12A8 

ElbowPlot(dabtramtodabtram) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
dabtramtodabtram <- FindNeighbors(dabtramtodabtram, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

dabtramtodabtram <- FindClusters(dabtramtodabtram, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 6906
Number of edges: 231004

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8325
Number of communities: 10
Elapsed time: 0 seconds

dabtramtodabtram <- RunUMAP(dabtramtodabtram, dims = 1:15)

10:43:18 UMAP embedding parameters a = 0.9922 b = 1.112
10:43:18 Read 6906 rows and found 15 numeric columns
10:43:18 Using Annoy for neighbor search, n_neighbors = 30
10:43:18 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:43:18 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f898b3add
10:43:18 Searching Annoy index using 1 thread, search_k = 3000
10:43:20 Annoy recall = 100%
10:43:20 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
10:43:21 Initializing from normalized Laplacian + noise (using irlba)
10:43:21 Commencing optimization for 500 epochs, with 286942 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:43:30 Optimization finished

DimPlot(dabtramtodabtram, reduction = 'umap', pt.size = 1)

# Get the scaled data from the dabtramtodabtram object
dabtramtodabtram_input_data <- GetAssayData(dabtramtodabtram, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
dabtramtodabtram_lin_pearson_list <- list()

for (i in fivecell_cDNA$DabTramtoDabTram){
  temp_pearson <- cor(dabtramtodabtram_input_data[,colnames(dabtramtodabtram)[dabtramtodabtram$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  dabtramtodabtram_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
dabtramtodabtram_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  dabtramtodabtram_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$DabTramtoDabTram){
    set.seed(j)
    num_cells <- length(dabtramtodabtram$Lineage[dabtramtodabtram$Lineage == i])
    temp_pearson <- cor(dabtramtodabtram_input_data[,sample(colnames(dabtramtodabtram), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    dabtramtodabtram_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - dabtramtocis

# Find the mean of the average pearson correlation per lineage
mean_pearson_dabtramtodabtram <- mean(unlist(lapply(dabtramtodabtram_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_dabtramtodabtram_sim <- sapply(1:length(dabtramtodabtram_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(dabtramtodabtram_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_dabtramtodabtram <- (mean_pearson_dabtramtodabtram-mean(means_pearson_dabtramtodabtram_sim))/sd(means_pearson_dabtramtodabtram_sim) # Z score comparing mean to simulations
pval_mean_pearson_dabtramtodabtram <- pnorm(z_mean_pearson_dabtramtodabtram, mean(means_pearson_dabtramtodabtram_sim), sd(means_pearson_dabtramtodabtram_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_dabtramtodabtram <- weighted.mean(unlist(lapply(dabtramtodabtram_lin_pearson_list, mean)),
unlist(lapply(dabtramtodabtram_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_dabtramtodabtram_sim <- sapply(1:length(dabtramtodabtram_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(dabtramtodabtram_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(dabtramtodabtram_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_dabtramtodabtram <- (weighted_mean_pearson_dabtramtodabtram-mean(weighted_means_pearson_dabtramtodabtram_sim))/sd(weighted_means_pearson_dabtramtodabtram_sim) # Z score comparing mean to simulations
pval_wmean_pearson_dabtramtodabtram <- pnorm(z_wmean_pearson_dabtramtodabtram, mean(weighted_means_pearson_dabtramtodabtram_sim), sd(weighted_means_pearson_dabtramtodabtram_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_dabtramtodabtram <- c()
for (i in 1:length(dabtramtodabtram_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(dabtramtodabtram_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_dabtramtodabtram <- cbind(wilcox_pval_dabtramtodabtram, wilcox.test(x = unlist(lapply(dabtramtodabtram_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(dabtramtodabtram, dabtramtodabtram_lin_pearson_list, dabtramtodabtram_lin_pearson_rand_list, z_mean_pearson_dabtramtodabtram, pval_mean_pearson_dabtramtodabtram, z_wmean_pearson_dabtramtodabtram, pval_wmean_pearson_dabtramtodabtram,  wilcox_pval_dabtramtodabtram, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/dabtramtodabtram_pearson_sim_results.RData')
rm(dabtramtodabtram, dabtramtodabtram_lin_pearson_list, dabtramtodabtram_lin_pearson_rand_list, dabtramtodabtram_input_data)

Significance testing of the dabtramtocis simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to dabtramtocis
dabtramtocis <- subset(all_data, idents = 'dabtramtocis') # Subset down to the dabtramtocis object
dabtramtocis <- NormalizeData(dabtramtocis)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtocis <- FindVariableFeatures(dabtramtocis, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtocis <- ScaleData(dabtramtocis)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

dabtramtocis <- RunPCA(dabtramtocis)

PC_ 1 
Positive:  HNRNPAB, MKI67, TUBA1B, TPM3, PRKDC, HMGB1, MT-ATP6, CFL1, ACTB, CENPF 
       YWHAB, CYCS, MT-CO3, H2AFZ, TPX2, SFPQ, CALU, PARP1, UBE2S, FASN 
       PFN1, TMPO, ASPM, CALR, MCM4, YWHAZ, ATAD2, RANBP1, STMN1, MYH10 
Negative:  GAS5, SAT1, MALAT1, SNHG7, EPB41L4A-AS1, TAF1D, GAPDH, HSPA1A, WSB1, NDRG1 
       BNIP3, SNHG12, HSPA1B, NRN1, DDIT4, SLC2A3, NT5C3A, CLK1, MYC, RGS1 
       RGS2, CREBRF, ZNF581, BBC3, POU3F1, ANKRD37, SH3BP4, RIPK4, APOD, HIST3H2A 
PC_ 2 
Positive:  MT-CO3, MT-CO2, MT-ATP6, MT-CYB, MT-ND3, MT-ND2, MT-ND1, MT-ND4, MARCKSL1, ARPC1B 
       CD63, LY6E, S100B, CCND1, FTH1, NOLC1, NDUFAF8, NDUFC2, HIPK2, PRKDC 
       FABP5, CFL1, NDUFB2, FASN, S100A11, METRN, SRM, PARP1, WWTR1, ATOX1 
Negative:  MKI67, CENPF, TOP2A, ASPM, TPX2, UBE2C, NUSAP1, HMGB2, PRC1, GTSE1 
       CENPE, PLK1, CCNB1, HMMR, KIF14, NUF2, KIF2C, AURKA, ARL6IP1, CENPA 
       DLGAP5, DEPDC1, SGO2, NEK2, ANLN, CEP55, CCNA2, CDCA8, BIRC5, PRR11 
PC_ 3 
Positive:  GAPDH, ACTG1, BNIP3, RPS8, FTL, SCD, EMP1, PGK1, DDIT4, RPL12 
       TPI1, MYC, GAS5, CSTB, SH3BP4, RPS6, TIMP3, NT5C3A, BHLHE40, SNHG7 
       PKM, UPP1, LGALS1, FTH1, NRN1, CTSD, TMSB10, NDRG1, LY6E, GYPC 
Negative:  MT-ATP6, MT-CO3, HELLS, ATAD2, MT-ND3, CLSPN, XRCC2, C21orf58, HIST1H1B, HIST1H1D 
       BRCA1, FAM111A, MT-CYB, MT-CO2, BRCA2, HMGA2-AS1, MALAT1, RRM2, FANCA, NEAT1 
       POLQ, PCNA, HIST2H2AC, OR2AT4, MCM4, POLD3, ORC6, AC058791.1, ATAD5, AUXG01000058.1 
PC_ 4 
Positive:  MT-ATP6, MT-CO3, MT-CYB, MT-ND3, MT-CO2, CCNB1, CENPE, CDC20, PLK1, PTTG1 
       JPT1, UBE2S, CDKN3, CCNB2, CD63, DYNLL1, HMMR, BIRC5, PTMS, AURKA 
       KIF14, ARL6IP1, CENPA, PRDX1, DLGAP5, PRR11, TPX2, MT-ND2, NEK2, ASPM 
Negative:  MCM4, HIST1H1B, ATAD2, CLSPN, RRM2, HELLS, MCM3, PCNA, HIST1H1D, PCLAF 
       DTL, GAPDH, BRCA1, POLD3, HIST1H4C, MALAT1, CENPU, HIST2H2AC, BHLHE40, BNIP3 
       FAM111A, E2F1, DDIT4, MCM7, CCNE2, MCM5, NDRG1, ENO2, ADM, CHAF1A 
PC_ 5 
Positive:  PRDX1, PSMA7, S100A6, TMSB10, ATOX1, FTL, MT2A, HSP90AA1, H2AFZ, CD63 
       FTH1, CHCHD2, CSTB, LGALS1, DBI, GSTO1, PHLDA2, SH3BGRL3, HSPE1, S100A11 
       ATP5MC3, CALM1, FABP5, POMP, HMGB1, NUDT1, ACTB, TMSB4X, VGF, NME1 
Negative:  NEAT1, MALAT1, GABPB1-AS1, BTG1, MT-ND2, MT-ND1, PDE3B, HPS4, SPTBN1, MT-ND6 
       MT-ND3, PIK3R3, MACF1, SAT1, NHLRC3, MT-CO2, MT-ND4, SNHG14, ZFP36L1, HIPK2 
       VMP1, CPM, ZNF106, FN1, MT-CYB, MT-CO3, AKAP12, SOX4, MYO10, MIR3142HG 

ElbowPlot(dabtramtocis) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
dabtramtocis <- FindNeighbors(dabtramtocis, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

dabtramtocis <- FindClusters(dabtramtocis, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 7070
Number of edges: 253207

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8557
Number of communities: 9
Elapsed time: 0 seconds

dabtramtocis <- RunUMAP(dabtramtocis, dims = 1:15)

10:46:18 UMAP embedding parameters a = 0.9922 b = 1.112
10:46:18 Read 7070 rows and found 15 numeric columns
10:46:18 Using Annoy for neighbor search, n_neighbors = 30
10:46:18 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:46:19 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f813c58810
10:46:19 Searching Annoy index using 1 thread, search_k = 3000
10:46:21 Annoy recall = 100%
10:46:21 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
10:46:22 Initializing from normalized Laplacian + noise (using irlba)
10:46:22 Commencing optimization for 500 epochs, with 300562 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:46:31 Optimization finished

DimPlot(dabtramtocis, reduction = 'umap', pt.size = 1)

# Get the scaled data from the dabtramtocis object
dabtramtocis_input_data <- GetAssayData(dabtramtocis, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
dabtramtocis_lin_pearson_list <- list()

for (i in fivecell_cDNA$DabTramtoCis){
  temp_pearson <- cor(dabtramtocis_input_data[,colnames(dabtramtocis)[dabtramtocis$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  dabtramtocis_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
dabtramtocis_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  dabtramtocis_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$DabTramtoCis){
    set.seed(j)
    num_cells <- length(dabtramtocis$Lineage[dabtramtocis$Lineage == i])
    temp_pearson <- cor(dabtramtocis_input_data[,sample(colnames(dabtramtocis), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    dabtramtocis_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - dabtramtococl2

# Find the mean of the average pearson correlation per lineage
mean_pearson_dabtramtocis <- mean(unlist(lapply(dabtramtocis_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_dabtramtocis_sim <- sapply(1:length(dabtramtocis_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(dabtramtocis_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_dabtramtocis <- (mean_pearson_dabtramtocis-mean(means_pearson_dabtramtocis_sim))/sd(means_pearson_dabtramtocis_sim) # Z score comparing mean to simulations
pval_mean_pearson_dabtramtocis <- pnorm(z_mean_pearson_dabtramtocis, mean(means_pearson_dabtramtocis_sim), sd(means_pearson_dabtramtocis_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_dabtramtocis <- weighted.mean(unlist(lapply(dabtramtocis_lin_pearson_list, mean)),
unlist(lapply(dabtramtocis_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_dabtramtocis_sim <- sapply(1:length(dabtramtocis_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(dabtramtocis_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(dabtramtocis_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_dabtramtocis <- (weighted_mean_pearson_dabtramtocis-mean(weighted_means_pearson_dabtramtocis_sim))/sd(weighted_means_pearson_dabtramtocis_sim) # Z score comparing mean to simulations
pval_wmean_pearson_dabtramtocis <- pnorm(z_wmean_pearson_dabtramtocis, mean(weighted_means_pearson_dabtramtocis_sim), sd(weighted_means_pearson_dabtramtocis_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_dabtramtocis <- c()
for (i in 1:length(dabtramtocis_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(dabtramtocis_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_dabtramtocis <- cbind(wilcox_pval_dabtramtocis, wilcox.test(x = unlist(lapply(dabtramtocis_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(dabtramtocis, dabtramtocis_lin_pearson_list, dabtramtocis_lin_pearson_rand_list, z_mean_pearson_dabtramtocis, pval_mean_pearson_dabtramtocis, z_wmean_pearson_dabtramtocis, pval_wmean_pearson_dabtramtocis,  wilcox_pval_dabtramtocis, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/dabtramtocis_pearson_sim_results.RData')
rm(dabtramtocis, dabtramtocis_lin_pearson_list, dabtramtocis_lin_pearson_rand_list, dabtramtocis_input_data)

Significance testing of the dabtramtococl2 simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to dabtram
dabtramtococl2 <- subset(all_data, idents = 'dabtramtococl2') # Subset down to the dabtram object
dabtramtococl2 <- NormalizeData(dabtramtococl2)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtococl2 <- FindVariableFeatures(dabtramtococl2, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

dabtramtococl2 <- ScaleData(dabtramtococl2)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

dabtramtococl2 <- RunPCA(dabtramtococl2)

PC_ 1 
Positive:  HIST1H4C, H2AFZ, TUBB, UBE2S, DTYMK, TUBA1B, GTSE1, UBE2C, C12orf75, CKS1B 
       TAGLN2, BIRC5, HMGB1, STMN1, TK1, TPX2, NME2, RANBP1, NUDT1, RPS5 
       JPT1, GYPC, ACTG1, EEF1D, TUBA1C, PRELID1, FOSL1, PHLDA2, CCDC85B, PTTG1 
Negative:  NEAT1, AC058791.1, MMP1, HMGA2-AS1, TTN, IGFN1, A2M, SNHG14, CBLB, S100B 
       CDH19, MMP3, SLC7A11-AS1, SOX10, LINC01705, CXCL8, CADPS, GPM6B, SULT1C2, PTGS2 
       MIR3142HG, MLANA, DCT, FBXO32, PRSS35, LINC00513, HSPA6, SLC5A3, AL139383.1, HSPA1B 
PC_ 2 
Positive:  MKI67, CENPF, ASPM, TOP2A, TPX2, TIMP3, PRC1, ANLN, BIRC5, RRM2 
       HJURP, IGFBP5, MT-ND3, VCL, STMN1, NUSAP1, AURKA, COL1A1, GTSE1, CCND1 
       COL6A1, KIF11, HIST1H1D, CENPE, CEP55, UBE2C, CCNB1, DEPDC1, CAV1, HMGB2 
Negative:  LINC00520, DDIT3, SAT1, GAS5, CSTB, AL118516.1, SNHG7, SNHG12, PDRG1, SNHG15 
       SQSTM1, ATP6V1F, MAP1LC3B, HMOX1, ATP6V0B, MT1X, ADRM1, ZFAND2A, AC003092.1, EIF5 
       ATF4, GADD45B, KLHL21, ODC1, SERTAD1, MED10, MED31, FBXO32, SNHG6, HSPA1A 
PC_ 3 
Positive:  CENPF, MKI67, TPX2, ASPM, TOP2A, ARL6IP1, NUSAP1, UBE2C, CCNB1, BIRC5 
       PRC1, PTTG1, AURKA, GTSE1, RRM2, UBE2S, HSP90AA1, CEP55, JPT1, HSPA1A 
       ANLN, KPNA2, KIF2C, CKS2, CENPE, CENPA, HMMR, GADD45B, AURKB, CDKN3 
Negative:  TIMP3, IGFBP7, CAV1, SFRP1, CALD1, GNG11, CTHRC1, TMEM158, IGFBP2, COL6A1 
       IGFBP5, COL6A2, SCG2, TMSB4X, NQO1, BIRC7, S100A13, SERPINE2, CADM1, CCPG1 
       CHCHD10, CST3, RPL34, CCN4, AKR1B1, UBE2E3, COL1A1, KYNU, MPC2, LOXL2 
PC_ 4 
Positive:  SERPINB2, TFPI2, RND3, TMSB4X, MMP1, IL24, SCG5, IER3, DKK1, EREG 
       AC003092.1, MMP3, LINC02376, CITED2, BASP1, IER2, ARL4C, KBTBD8, TNFRSF12A, ERRFI1 
       JUN, GDF15, LINC00973, DDIT3, CXCL8, IGFN1, IL1B, PPP1R15A, GADD45A, STC1 
Negative:  MLANA, S100B, SOX10, MIA, PMEL, TYR, GAS7, GPM6B, FXYD3, RAMP1 
       MFSD12, PLP1, NFATC2, DCT, SLC1A4, LHFPL3-AS1, SLAMF9, CDH19, CITED1, SORBS2 
       RAB38, ERBB3, TNS1, AL035541.1, SOX6, ZNF704, AKAP12, UCN2, COL9A3, A2M 
PC_ 5 
Positive:  ARL4C, PRNP, FRMD4A, MME, CADM1, SRGN, IGFN1, TMSB4X, ARHGAP18, SIRPB1 
       SFRP1, SERPINB2, SPOCK1, BIRC7, GPNMB, MBP, MAP4K4, TNIK, LINC02376, SPP1 
       TMEM158, HAS2, TGFBR1, DNER, HCFC1R1, NR2F2, BACE2, CDC42EP3, PMP22, CAPG 
Negative:  PLAT, TRPA1, CTSC, IGFBP5, SLC14A1, TNFRSF11B, MMD, BOC, CBLN2, TMEM100 
       COLEC10, KYNU, TBX2, F3, C11orf96, FAM20C, FOXF1, BMP2, CDCP1, FENDRR 
       TMX4, HSD17B2, PCDH18, ACSL4, BAMBI, ISG15, PHACTR2, AKR1B1, SIX1, AL353653.1 

ElbowPlot(dabtramtococl2) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
dabtramtococl2 <- FindNeighbors(dabtramtococl2, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

dabtramtococl2 <- FindClusters(dabtramtococl2, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 10184
Number of edges: 315953

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8518
Number of communities: 11
Elapsed time: 1 seconds

dabtramtococl2 <- RunUMAP(dabtramtococl2, dims = 1:15)

10:49:12 UMAP embedding parameters a = 0.9922 b = 1.112
10:49:12 Read 10184 rows and found 15 numeric columns
10:49:12 Using Annoy for neighbor search, n_neighbors = 30
10:49:12 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:49:13 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f8d67dc94
10:49:13 Searching Annoy index using 1 thread, search_k = 3000
10:49:16 Annoy recall = 100%
10:49:16 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
10:49:17 Initializing from normalized Laplacian + noise (using irlba)
10:49:17 Commencing optimization for 200 epochs, with 431274 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:49:22 Optimization finished

DimPlot(dabtramtococl2, reduction = 'umap', pt.size = 1)

# Get the scaled data from the dabtramtococl2 object
dabtramtococl2_input_data <- GetAssayData(dabtramtococl2, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
dabtramtococl2_lin_pearson_list <- list()

for (i in fivecell_cDNA$DabTramtoCoCl2){
  temp_pearson <- cor(dabtramtococl2_input_data[,colnames(dabtramtococl2)[dabtramtococl2$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  dabtramtococl2_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
dabtramtococl2_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  dabtramtococl2_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$DabTramtoCoCl2){
    set.seed(j)
    num_cells <- length(dabtramtococl2$Lineage[dabtramtococl2$Lineage == i])
    temp_pearson <- cor(dabtramtococl2_input_data[,sample(colnames(dabtramtococl2), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    dabtramtococl2_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cocl2

# Find the mean of the average pearson correlation per lineage
mean_pearson_dabtramtococl2 <- mean(unlist(lapply(dabtramtococl2_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_dabtramtococl2_sim <- sapply(1:length(dabtramtococl2_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(dabtramtococl2_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_dabtramtococl2 <- (mean_pearson_dabtramtococl2-mean(means_pearson_dabtramtococl2_sim))/sd(means_pearson_dabtramtococl2_sim) # Z score comparing mean to simulations
pval_mean_pearson_dabtramtococl2 <- pnorm(z_mean_pearson_dabtramtococl2, mean(means_pearson_dabtramtococl2_sim), sd(means_pearson_dabtramtococl2_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_dabtramtococl2 <- weighted.mean(unlist(lapply(dabtramtococl2_lin_pearson_list, mean)),
unlist(lapply(dabtramtococl2_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_dabtramtococl2_sim <- sapply(1:length(dabtramtococl2_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(dabtramtococl2_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(dabtramtococl2_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_dabtramtococl2 <- (weighted_mean_pearson_dabtramtococl2-mean(weighted_means_pearson_dabtramtococl2_sim))/sd(weighted_means_pearson_dabtramtococl2_sim) # Z score comparing mean to simulations
pval_wmean_pearson_dabtramtococl2 <- pnorm(z_wmean_pearson_dabtramtococl2, mean(weighted_means_pearson_dabtramtococl2_sim), sd(weighted_means_pearson_dabtramtococl2_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_dabtramtococl2 <- c()
for (i in 1:length(dabtramtococl2_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(dabtramtococl2_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_dabtramtococl2 <- cbind(wilcox_pval_dabtramtococl2, wilcox.test(x = unlist(lapply(dabtramtococl2_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(dabtramtococl2, dabtramtococl2_lin_pearson_list, dabtramtococl2_lin_pearson_rand_list, z_mean_pearson_dabtramtococl2, pval_mean_pearson_dabtramtococl2, z_wmean_pearson_dabtramtococl2, pval_wmean_pearson_dabtramtococl2,  wilcox_pval_dabtramtococl2, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/dabtramtococl2_pearson_sim_results.RData')
rm(dabtramtococl2, dabtramtococl2_lin_pearson_list, dabtramtococl2_lin_pearson_rand_list, dabtramtococl2_input_data)

Significance testing of the cocl2 simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cocl2
cocl2 <- subset(all_data, idents = 'cocl2') # Subset down to the cocl2 object
cocl2 <- NormalizeData(cocl2)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2 <- FindVariableFeatures(cocl2, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2 <- ScaleData(cocl2)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cocl2 <- RunPCA(cocl2)

PC_ 1 
Positive:  MLANA, PMEL, H2AFZ, CAPG, MT-CYB, MITF, STMN1, MKI67, HMGB1, LGALS3 
       CHCHD6, TPX2, CENPF, FXYD3, MT-ND2, CDK2, RGS10, TOP2A, FRMD4B, ATOX1 
       ASPM, PIK3R3, ACTG1, NUSAP1, QPCT, TIMM50, TSPAN10, BIRC5, RPL12, AURKB 
Negative:  MT2A, UBC, MT1X, TNFRSF12A, JUN, FN1, CLEC2B, GNG11, TMEM158, PPP1R15A 
       FTL, RGS2, ANXA1, AC003092.1, TFPI2, ANGPTL4, DKK1, LINC00973, BASP1, PHLDA1 
       PRNP, CITED2, DDIT3, LUCAT1, EIF1, GAS5, MALAT1, IER2, VGF, FTH1 
PC_ 2 
Positive:  SAT1, PMEL, ERBB3, PDE3B, RAMP1, APOE, RPL23, GAS5, TYR, TDRD3 
       NFATC2, MIA, RGS10, LHFPL3-AS1, EPB41L4A-AS1, RAB38, MLANA, C11orf96, ZFAS1, TFAP2A 
       SLC44A1, MYC, CBLB, BCAN, NEAT1, TRMT9B, PIK3R3, EDNRB, STK32A, CADPS 
Negative:  UBE2S, TUBA1B, TUBB, MKI67, CENPF, UBE2C, CKS1B, HMGB2, TPX2, TOP2A 
       BIRC5, NUSAP1, GTSE1, ACTB, RRM2, TK1, DTYMK, PRC1, ANLN, FOSL1 
       CALM2, UBE2T, SMC4, TUBB4B, TMSB4X, HIST1H4C, PCLAF, CDK1, C12orf75, CKS2 
PC_ 3 
Positive:  HSPA1A, GADD45B, SNHG12, DDIT3, SNHG7, HSPA1B, ZFAS1, SNHG1, GAS5, KBTBD8 
       CKS2, SERTAD1, PDRG1, SNHG15, ZFAND2A, ATF3, MAPKAPK5-AS1, HSPH1, PCLO, DNAJB1 
       ATF4, PPP1R15A, HIST1H4H, MIR4300HG, HSP90AA1, IER2, PIGL, IGFL2-AS1, IFRD1, LINC00520 
Negative:  SERPINE2, S100A6, TIMP3, CST3, TIMP1, GNAS, S100A13, LY6E, SH3BGRL3, FN1 
       CCND1, TMEM158, CAV1, PRSS35, MIA, VKORC1, S100A1, CADM1, SLC20A1, PRNP 
       PRSS23, ME1, CTHRC1, FGFBP2, DCBLD2, CDKN2A, P4HB, CALD1, PRRX1, SPARC 
PC_ 4 
Positive:  FN1, CD44, THBS1, MALAT1, SRGN, DST, MAP4K4, AHNAK, MT-ND6, IGFBP7 
       F2R, IGFN1, LMO7, NEAT1, COL6A2, ASPM, MMP2, SERPINB2, EREG, SCG2 
       ARL4C, IL6ST, PAG1, GABPB1-AS1, HIST1H1D, CBLB, TXNRD1, MYOF, TFPI2, F3 
Negative:  NDUFC2, PSMA7, CSTB, ATOX1, SEC61G, NDUFA4, UPP1, FAM162A, PRDX1, LGALS3 
       CALM1, BIRC7, ATP6V0B, MAP1LC3B, TXNDC17, BNIP3, C4orf3, RAB5IF, ATP5MC3, VKORC1 
       SEC11C, QPCT, UCN2, DBI, GSTP1, TMSB10, PGK1, MIA, S100A11, GNAS 
PC_ 5 
Positive:  RPL12, LGALS1, FTL, SRGN, FTH1, TFPI2, SERPINB2, TMSB4X, TMSB10, TXN 
       RPL23, SCG5, C12orf75, CHCHD10, LINC02376, MYEOV, BASP1, ATP5MC3, MMP1, DBI 
       GYPC, ACTG1, S100A11, F3, CSF1, IGFBP7, RGS10, BOLA3, VEGFC, KYNU 
Negative:  MALAT1, NEAT1, SLC5A3, VMP1, DST, FAT1, SERPINE2, ZFYVE16, NFKBIZ, TXNIP 
       GABPB1-AS1, SGK1, CADM1, SLC2A3, USP53, APOD, COL9A3, SLC20A1, HSP90B1, CDH19 
       MFGE8, PRSS35, GAS7, SNHG14, CBLB, LEF1, GLS, NORAD, FGFBP2, SPARC 

ElbowPlot(cocl2) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cocl2 <- FindNeighbors(cocl2, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cocl2 <- FindClusters(cocl2, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 4823
Number of edges: 164423

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8480
Number of communities: 11
Elapsed time: 0 seconds

cocl2 <- RunUMAP(cocl2, dims = 1:15)

10:52:30 UMAP embedding parameters a = 0.9922 b = 1.112
10:52:30 Read 4823 rows and found 15 numeric columns
10:52:30 Using Annoy for neighbor search, n_neighbors = 30
10:52:30 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:52:31 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f81dc27f6c
10:52:31 Searching Annoy index using 1 thread, search_k = 3000
10:52:32 Annoy recall = 100%
10:52:32 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
10:52:33 Initializing from normalized Laplacian + noise (using irlba)
10:52:33 Commencing optimization for 500 epochs, with 201272 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:52:40 Optimization finished

DimPlot(cocl2, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cocl2 object
cocl2_input_data <- GetAssayData(cocl2, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cocl2_lin_pearson_list <- list()

for (i in fivecell_cDNA$CoCl2){
  temp_pearson <- cor(cocl2_input_data[,colnames(cocl2)[cocl2$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cocl2_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cocl2_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cocl2_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CoCl2){
    set.seed(j)
    num_cells <- length(cocl2$Lineage[cocl2$Lineage == i])
    temp_pearson <- cor(cocl2_input_data[,sample(colnames(cocl2), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cocl2_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cocl2tococl2

# Find the mean of the average pearson correlation per lineage
mean_pearson_cocl2 <- mean(unlist(lapply(cocl2_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cocl2_sim <- sapply(1:length(cocl2_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cocl2_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cocl2 <- (mean_pearson_cocl2-mean(means_pearson_cocl2_sim))/sd(means_pearson_cocl2_sim) # Z score comparing mean to simulations
pval_mean_pearson_cocl2 <- pnorm(z_mean_pearson_cocl2, mean(means_pearson_cocl2_sim), sd(means_pearson_cocl2_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cocl2 <- weighted.mean(unlist(lapply(cocl2_lin_pearson_list, mean)),
unlist(lapply(cocl2_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cocl2_sim <- sapply(1:length(cocl2_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cocl2_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cocl2_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cocl2 <- (weighted_mean_pearson_cocl2-mean(weighted_means_pearson_cocl2_sim))/sd(weighted_means_pearson_cocl2_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cocl2 <- pnorm(z_wmean_pearson_cocl2, mean(weighted_means_pearson_cocl2_sim), sd(weighted_means_pearson_cocl2_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cocl2 <- c()
for (i in 1:length(cocl2_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cocl2_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cocl2 <- cbind(wilcox_pval_cocl2, wilcox.test(x = unlist(lapply(cocl2_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(cocl2, cocl2_lin_pearson_list, cocl2_lin_pearson_rand_list, z_mean_pearson_cocl2, pval_mean_pearson_cocl2, z_wmean_pearson_cocl2, pval_wmean_pearson_cocl2,  wilcox_pval_cocl2, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cocl2_pearson_sim_results.RData')
rm(cocl2, cocl2_lin_pearson_list, cocl2_lin_pearson_rand_list, cocl2_input_data)

Significance testing of the cocl2tococl2 simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cocl2tococl2
cocl2tococl2 <- subset(all_data, idents = 'cocl2tococl2') # Subset down to the cocl2tococl2 object
cocl2tococl2 <- NormalizeData(cocl2tococl2)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2tococl2 <- FindVariableFeatures(cocl2tococl2, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2tococl2 <- ScaleData(cocl2tococl2)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cocl2tococl2 <- RunPCA(cocl2tococl2)

PC_ 1 
Positive:  RPS8, RPL10, RPS2, RPS3, RPS14, GAPDH, RPL28, RPL41, FTH1, RPS5 
       RPS12, RPS16, RPL26, EEF1A1, RPL11, RPS15A, RPL12, FTL, RPS4X, RPL29 
       RPS11, RPL7A, RPL15, RPL8, RPL17, RPS6, RPS3A, RPS23, RPS13, RPL19 
Negative:  MALAT1, NEAT1, WSB1, PLAC4, MT-ND6, NFKBIZ, LINC00488, SNHG14, PAXBP1, MT-ATP6 
       HPS4, HMGA2-AS1, SLC5A3, HSPA1B, NKTR, HSPA1A, ANKRD11, PRR26, AL162426.1, POLR2J3 
       CLCN7, CCDC144A, VMP1, FTX, AC005632.3, AC003681.1, ATP1A1-AS1, AC058791.1, PLCG2, N4BP2L2 
PC_ 2 
Positive:  NEAT1, MKI67, CENPF, ASPM, MALAT1, ANKRD11, GTSE1, CIT, POLR2J3, FRMD4B 
       TPX2, SLC7A11, PRKDC, PRC1, TOP2A, SFPQ, NAV2, NUSAP1, ZFYVE16, UBE2C 
       MITF, HPS4, WSB1, HSP90B1, RAD21, MAT2A, SLC20A1, SLC25A37, CALR, AHNAK 
Negative:  RPL26, RPS12, RPL41, RPS23, FTH1, RPL10, RPL11, RPS16, RPS14, RPS2 
       FTL, RPS11, RPS3A, RPL34, RPS3, RPL28, RPL29, RPS4X, RPL17, RPL7A 
       RPS15A, EEF1A1, RPL35A, RPL8, RPL15, RPL23, RPS13, RPS5, GAPDH, RPS6 
PC_ 3 
Positive:  NEAT1, GABPB1-AS1, SAT1, SNHG14, RPS8, ZKSCAN1, PRAME, LHFPL3-AS1, POLR2J3, ERBB3 
       TRIM25, PLEC, FAT1, CADPS, BTG1, CCNI, CBLB, N4BP2L2, TXNIP, TFAP2A 
       SLC25A37, DANCR, SLC5A3, ZNF106, SLC2A3, SCD, MT-ND3, SEZ6L2, CSPG4, PDE3B 
Negative:  CENPF, MKI67, TPX2, UBE2C, TOP2A, ASPM, HMGB2, AURKA, NUSAP1, CCNB1 
       PRC1, UBE2S, GTSE1, BIRC5, KIF2C, KPNA2, PLK1, PTTG1, CEP55, HMMR 
       CKAP2, CKS2, ARL6IP1, TUBB4B, CCNA2, CENPE, CDC20, PRR11, TUBA1B, DLGAP5 
PC_ 4 
Positive:  MT-ATP6, MT-ND3, GABPB1-AS1, LHFPL3-AS1, MT-ND6, PIK3R3, LINC00488, MIR3142HG, SLC16A1-AS1, ZKSCAN1 
       N4BP2L2, HPS4, ASPM, FRMD4B, FTX, NFIA, HIPK2, SEMA6A, ERBB3, TSPAN10 
       PDE3B, AC063923.2, LRRC75A, NPAT, TRIM73, PRKDC, RPS3, AC008170.1, HELLS, OR2AT4 
Negative:  MT2A, IER3, HSPA1A, DNAJB1, NDRG1, ADM, HSPA1B, S100A6, TMEM158, TIMP1 
       HSP90AA1, MT1X, RGS2, SLC5A3, DDIT4, SERPINE2, FN1, HILPDA, SLC20A1, SH3BGRL3 
       GAPDH, TIMP3, VGF, FTL, TMSB10, PPP1R15A, PRSS35, RND3, PLIN2, SPP1 
PC_ 5 
Positive:  HSPA1A, DNAJB1, HSPA1B, SAT1, HSP90AA1, RGS1, NFKBIZ, HPS4, PPP1R15A, KLF6 
       SLC2A3, CHORDC1, GAS5, SNHG7, IRS2, BBC3, ID3, C11orf96, FOS, ID2 
       RPL10, NKTR, MALAT1, RGS10, TECR, YPEL2, ZFP36L1, BTG1, EPB41L4A-AS1, NR4A1 
Negative:  MT-ATP6, MT-ND3, S100A6, MT-ND6, FN1, SH3BGRL3, DCBLD2, TMSB10, TMSB4X, HNRNPAB 
       FTL, VGF, TMEM158, FTH1, STC1, FABP5, TXNRD1, PRKDC, PHLDA2, GABPB1-AS1 
       TPM3, CFL1, TNFRSF12A, PFN1, PKM, ATOX1, PRDX1, CALM1, ATAD2, YWHAB 

ElbowPlot(cocl2tococl2) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cocl2tococl2 <- FindNeighbors(cocl2tococl2, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cocl2tococl2 <- FindClusters(cocl2tococl2, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 13951
Number of edges: 400713

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8156
Number of communities: 8
Elapsed time: 2 seconds

cocl2tococl2 <- RunUMAP(cocl2tococl2, dims = 1:15)

10:55:09 UMAP embedding parameters a = 0.9922 b = 1.112
10:55:09 Read 13951 rows and found 15 numeric columns
10:55:09 Using Annoy for neighbor search, n_neighbors = 30
10:55:09 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:55:10 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f84c3b9ab7
10:55:10 Searching Annoy index using 1 thread, search_k = 3000
10:55:14 Annoy recall = 100%
10:55:15 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
10:55:15 Initializing from normalized Laplacian + noise (using irlba)
10:55:15 Commencing optimization for 200 epochs, with 602270 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
10:55:23 Optimization finished

DimPlot(cocl2tococl2, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cocl2tococl2 object
cocl2tococl2_input_data <- GetAssayData(cocl2tococl2, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cocl2tococl2_lin_pearson_list <- list()

for (i in fivecell_cDNA$CoCl2toCoCl2){
  temp_pearson <- cor(cocl2tococl2_input_data[,colnames(cocl2tococl2)[cocl2tococl2$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cocl2tococl2_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cocl2tococl2_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cocl2tococl2_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CoCl2toCoCl2){
    set.seed(j)
    num_cells <- length(cocl2tococl2$Lineage[cocl2tococl2$Lineage == i])
    temp_pearson <- cor(cocl2tococl2_input_data[,sample(colnames(cocl2tococl2), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cocl2tococl2_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cocl2tocis

# Find the mean of the average pearson correlation per lineage
mean_pearson_cocl2tococl2 <- mean(unlist(lapply(cocl2tococl2_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cocl2tococl2_sim <- sapply(1:length(cocl2tococl2_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cocl2tococl2_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cocl2tococl2 <- (mean_pearson_cocl2tococl2-mean(means_pearson_cocl2tococl2_sim))/sd(means_pearson_cocl2tococl2_sim) # Z score comparing mean to simulations
pval_mean_pearson_cocl2tococl2 <- pnorm(z_mean_pearson_cocl2tococl2, mean(means_pearson_cocl2tococl2_sim), sd(means_pearson_cocl2tococl2_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cocl2tococl2 <- weighted.mean(unlist(lapply(cocl2tococl2_lin_pearson_list, mean)),
unlist(lapply(cocl2tococl2_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cocl2tococl2_sim <- sapply(1:length(cocl2tococl2_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cocl2tococl2_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cocl2tococl2_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cocl2tococl2 <- (weighted_mean_pearson_cocl2tococl2-mean(weighted_means_pearson_cocl2tococl2_sim))/sd(weighted_means_pearson_cocl2tococl2_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cocl2tococl2 <- pnorm(z_wmean_pearson_cocl2tococl2, mean(weighted_means_pearson_cocl2tococl2_sim), sd(weighted_means_pearson_cocl2tococl2_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cocl2tococl2 <- c()
for (i in 1:length(cocl2tococl2_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cocl2tococl2_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cocl2tococl2 <- cbind(wilcox_pval_cocl2tococl2, wilcox.test(x = unlist(lapply(cocl2tococl2_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2tococl2_lin_pearson_list,  :
  cannot compute exact p-value with ties

# Save outputs
save(cocl2tococl2, cocl2tococl2_lin_pearson_list, cocl2tococl2_lin_pearson_rand_list, z_mean_pearson_cocl2tococl2, pval_mean_pearson_cocl2tococl2, z_wmean_pearson_cocl2tococl2, pval_wmean_pearson_cocl2tococl2,  wilcox_pval_cocl2tococl2, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cocl2tococl2_pearson_sim_results.RData')
rm(cocl2tococl2, cocl2tococl2_lin_pearson_list, cocl2tococl2_lin_pearson_rand_list, cocl2tococl2_input_data)

Significance testing of the cocl2tocis simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to dabtram
cocl2tocis <- subset(all_data, idents = 'cocl2tocis') # Subset down to the cocl2 object
cocl2tocis <- NormalizeData(cocl2tocis)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2tocis <- FindVariableFeatures(cocl2tocis, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2tocis <- ScaleData(cocl2tocis)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cocl2tocis <- RunPCA(cocl2tocis)

PC_ 1 
Positive:  MALAT1, SAT1, GAS5, RPL26, HSPA1A, SNHG7, PMEL, NEAT1, HSPA1B, NDRG1 
       RPS3, CREBRF, EPB41L4A-AS1, RPL7A, LINC01531, CLK1, SLC2A3, RGS2, APOD, NT5C3A 
       NUPR1, MYC, HIST3H2A, NHLRC3, PNRC1, SNHG14, RGS1, SLC5A3, DNAJB1, SNHG29 
Negative:  MKI67, ACTB, TUBA1B, MT-ATP6, HNRNPAB, PRKDC, CFL1, HMGB1, H2AFZ, TPX2 
       PARP1, CENPF, HIST1H4C, PFN1, CALR, DYNLL1, TMPO, YWHAB, HNRNPD, UBE2S 
       TPM3, DTYMK, NME1, YWHAZ, TUBB, CYCS, JPT1, PKM, AP2S1, ARPC1B 
PC_ 2 
Positive:  FTL, LY6E, FTH1, MT-ND3, MT-CYB, MT-ATP6, MT-CO2, ARPC1B, RPS8, SERF2 
       PKM, RPL12, TMSB10, SRM, S100A1, MT-ND1, NME4, MT-ND2, S100B, TPI1 
       MARCKSL1, CYBA, CHCHD10, ATP5F1E, CFL1, METRN, SH3BGRL3, DBI, AP2S1, POLR2F 
Negative:  ASPM, CENPF, MKI67, TPX2, TOP2A, NUSAP1, UBE2C, CENPE, AURKA, GTSE1 
       CCNB1, HMGB2, PRC1, KIF14, DEPDC1, ARL6IP1, SGO2, HMMR, PLK1, KIF2C 
       CENPA, TUBB4B, ANLN, NEK2, KPNA2, CCNA2, KIF23, NUF2, CEP55, DLGAP5 
PC_ 3 
Positive:  GAPDH, BNIP3, ACTG1, RPS8, CSTB, SNHG29, FTL, PGK1, BTG1, RPS3 
       MIF, DDIT4, LDHA, RPL12, GAS5, RPL26, SCD, MLANA, MYC, EMP1 
       TPI1, SNHG7, NT5C3A, SLAMF9, SH3BP4, S100A10, RGS10, ENO2, RPL29, FAM162A 
Negative:  ATAD2, MCM4, MT-ATP6, CLSPN, HELLS, PCNA, XRCC2, BRCA1, HIST1H1B, MT-ND6 
       DTL, MCM3, MCM6, HIST1H1D, MT-CO2, GINS2, CCNE2, BRCA2, RRM2, E2F1 
       POLD3, HIST2H2AC, MT-CYB, CHAF1A, MCM10, FANCA, CDC6, MCM7, FEN1, MCM5 
PC_ 4 
Positive:  NEAT1, MALAT1, FN1, NDRG1, DDIT4, GAPDH, BNIP3, SLC5A3, SLC2A3, RGS2 
       SERPINE2, ADM, NT5C3A, BHLHE40, PPP1R15A, TNS1, A2M, ZFP36L1, ATAD2, CAV1 
       TIMP3, HIST1H1B, ENO2, NRN1, BIRC7, FGFBP2, CADM1, RND3, COL6A1, CCDC85B 
Negative:  PRDX1, CD63, ATOX1, AKR1A1, PSMA7, H2AFZ, MT-ATP6, DYNLL1, JPT1, PTTG1 
       UBE2S, FABP5, HMGB1, TXN, NDUFC2, DBI, CDKN3, POMP, CCNB1, CKS2 
       ARPC1B, HSP90AA1, CALM1, PMEL, PFN1, MLANA, EIF5A, CHCHD6, DCT, HMGB3 
PC_ 5 
Positive:  FN1, AHNAK, IER3, MT-CO2, MT-ATP6, MT-ND1, SERPINE2, MT-CYB, MT-ND2, CCND1 
       SLC20A1, MT-ND3, MT-ND6, SORBS2, COL6A1, IGFBP7, CAV1, LGALS1, S100A4, NFATC2 
       ANXA2, PRRX1, SPP1, PRNP, ATP2B1, CALU, PLEC, ANXA1, PCOLCE, CDK6 
Negative:  RPL26, RPS3, HIST1H1B, HIST1H4C, PMEL, RPL7A, RPL29, HIST1H1D, MLANA, HIST2H2AC 
       SNHG29, RPS8, RGS10, RPL12, RRM2, PCLAF, HIST1H3D, HIST1H1E, ATAD2, AURKB 
       GAPDH, HIST1H1C, HIST1H2AH, CLSPN, HELLS, BRCA1, FAM111A, GAS5, HIST1H1A, CSTB 

ElbowPlot(cocl2tocis) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cocl2tocis <- FindNeighbors(cocl2tocis, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cocl2tocis <- FindClusters(cocl2tocis, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 6459
Number of edges: 229291

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8415
Number of communities: 9
Elapsed time: 0 seconds

cocl2tocis <- RunUMAP(cocl2tocis, dims = 1:15)

15:37:19 UMAP embedding parameters a = 0.9922 b = 1.112
15:37:19 Read 6459 rows and found 15 numeric columns
15:37:19 Using Annoy for neighbor search, n_neighbors = 30
15:37:19 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:37:20 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f83ce09dfb
15:37:20 Searching Annoy index using 1 thread, search_k = 3000
15:37:22 Annoy recall = 100%
15:37:28 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:37:29 Initializing from normalized Laplacian + noise (using irlba)
15:37:29 Commencing optimization for 500 epochs, with 272244 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:37:38 Optimization finished

DimPlot(cocl2tocis, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cocl2tocis object
cocl2tocis_input_data <- GetAssayData(cocl2tocis, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cocl2tocis_lin_pearson_list <- list()

for (i in fivecell_cDNA$CoCl2toCis){
  temp_pearson <- cor(cocl2tocis_input_data[,colnames(cocl2tocis)[cocl2tocis$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cocl2tocis_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cocl2tocis_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cocl2tocis_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CoCl2toCis){
    set.seed(j)
    num_cells <- length(cocl2tocis$Lineage[cocl2tocis$Lineage == i])
    temp_pearson <- cor(cocl2tocis_input_data[,sample(colnames(cocl2tocis), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cocl2tocis_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cocl2todabtram

# Find the mean of the average pearson correlation per lineage
mean_pearson_cocl2tocis <- mean(unlist(lapply(cocl2tocis_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cocl2tocis_sim <- sapply(1:length(cocl2tocis_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cocl2tocis_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cocl2tocis <- (mean_pearson_cocl2tocis-mean(means_pearson_cocl2tocis_sim))/sd(means_pearson_cocl2tocis_sim) # Z score comparing mean to simulations
pval_mean_pearson_cocl2tocis <- pnorm(z_mean_pearson_cocl2tocis, mean(means_pearson_cocl2tocis_sim), sd(means_pearson_cocl2tocis_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cocl2tocis <- weighted.mean(unlist(lapply(cocl2tocis_lin_pearson_list, mean)),
unlist(lapply(cocl2tocis_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cocl2tocis_sim <- sapply(1:length(cocl2tocis_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cocl2tocis_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cocl2tocis_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cocl2tocis <- (weighted_mean_pearson_cocl2tocis-mean(weighted_means_pearson_cocl2tocis_sim))/sd(weighted_means_pearson_cocl2tocis_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cocl2tocis <- pnorm(z_wmean_pearson_cocl2tocis, mean(weighted_means_pearson_cocl2tocis_sim), sd(weighted_means_pearson_cocl2tocis_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cocl2tocis <- c()
for (i in 1:length(cocl2tocis_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cocl2tocis_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cocl2tocis <- cbind(wilcox_pval_cocl2tocis, wilcox.test(x = unlist(lapply(cocl2tocis_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(cocl2tocis, cocl2tocis_lin_pearson_list, cocl2tocis_lin_pearson_rand_list, z_mean_pearson_cocl2tocis, pval_mean_pearson_cocl2tocis, z_wmean_pearson_cocl2tocis, pval_wmean_pearson_cocl2tocis,  wilcox_pval_cocl2tocis, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cocl2tocis_pearson_sim_results.RData')
rm(cocl2tocis, cocl2tocis_lin_pearson_list, cocl2tocis_lin_pearson_rand_lis, cocl2tocis_input_data)

Warning in rm(cocl2tocis, cocl2tocis_lin_pearson_list, cocl2tocis_lin_pearson_rand_lis,  :
  object 'cocl2tocis_lin_pearson_rand_lis' not found

Significance testing of the cocl2todabtram simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cocl2todabtram
cocl2todabtram <- subset(all_data, idents = 'cocl2todabtram') # Subset down to the cocl2todabtram object
cocl2todabtram <- NormalizeData(cocl2todabtram)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2todabtram <- FindVariableFeatures(cocl2todabtram, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cocl2todabtram <- ScaleData(cocl2todabtram)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cocl2todabtram <- RunPCA(cocl2todabtram)

PC_ 1 
Positive:  GAS7, SOX10, COL9A3, AKAP12, S100B, NFATC2, SOX6, PHACTR1, IRS2, TFAP2A 
       EPB41L3, OLFML2A, GPM6B, HSPG2, TSC22D1, EVI5, SEMA3B, PLP1, CRYAB, SCARB1 
       ERBB3, PDZRN3, LIMA1, CBLB, SOX4, PALLD, PMP22, AL139383.1, NCAM1, C1orf198 
Negative:  BASP1, CAV1, PHLDA1, COL1A1, COL6A2, GYPC, CCND1, IL6ST, SFRP1, TXNRD1 
       COL6A1, CDC25B, SLC12A8, ANXA1, ARHGAP18, PCLAF, C12orf75, TK1, CEMIP, VEGFC 
       CYTOR, TXN, PERP, ATP2B1, EGFR, ANLN, MKI67, CAPG, MAP1B, TFPI2 
PC_ 2 
Positive:  DAB2, SLC12A8, NUPR1, LGALS3, ARL4C, CTSK, ITGA11, PLXNC1, IL6ST, SGK1 
       BIRC7, ITGB8, CHCHD10, PDGFRB, GPNMB, TXNRD1, OLFML2B, LINC00968, COL12A1, C1S 
       C1R, CEBPB, DEPTOR, BAMBI, NABP1, COL14A1, FTH1, SPON2, MTSS1, ITGAV 
Negative:  MKI67, CENPF, ANLN, GTSE1, TPX2, CEP55, ASPM, PRC1, NUSAP1, TOP2A 
       UBE2C, BIRC5, NUF2, UBE2S, H2AFX, KIF11, CDCA2, HMGB2, DEPDC1, NCAPG 
       CCNA2, CENPA, CENPE, RRM2, HMMR, KIF23, CCNB1, STMN1, KIF2C, LMNB1 
PC_ 3 
Positive:  LOXL2, FN1, THBS1, SPARC, CALD1, COL5A2, MFGE8, HSPG2, FSTL1, TIMP3 
       CSRP2, MMP2, FGFR1, S100A6, LMO7, SERPINE2, CDC42EP3, ALCAM, PRNP, IFI6 
       MCAM, NUAK1, GREM2, PALLD, ARID5B, EVI5, RAPH1, OLFML2A, COL6A1, CTHRC1 
Negative:  PMEL, MLANA, CAPN3, LINC00520, TDRD3, LGALS3, BHLHE41, ID2, SULT1C2, AP1S2 
       AC004988.1, CITED1, ADGRG1, STK32A, CDK2, RAB38, TYR, BAAT, FBXO32, ATP6V0D2 
       CHCHD6, GAS5, TRPM1, TRIM63, ABCB5, BCL2A1, EDNRB, AL162457.2, SNHG12, SNHG7 
PC_ 4 
Positive:  GCLM, VCAM1, SLC7A11, FOXF1, KYNU, COL14A1, PKDCC, PTGR1, RHOBTB3, CTSC 
       CASP1, PGD, TRIM16L, ASPH, C1R, FTH1, MGST1, PCDH18, SQSTM1, TBX2 
       CEBPD, IGFBP5, FENDRR, SRXN1, LINC01914, GSTM3, HHEX, SDCBP, BAMBI, SLC3A2 
Negative:  ITGA3, PRSS23, CPA4, SERPINE1, FRMD4A, TNIK, GCNT1, TIMP3, PHLDB2, C12orf75 
       ABI3BP, DKK1, PLP2, BDNF, EVI2A, AXL, FGF5, ATP1B1, F2R, PDGFC 
       SPOCD1, GRAMD2B, MT2A, RPS27L, AC092807.3, TPST2, MEGF6, LMO7, SRGN, VGF 
PC_ 5 
Positive:  HELLS, ATAD2, DTL, MCM3, CCNE2, CLSPN, BRCA1, E2F1, MCM4, FANCA 
       POLD3, GINS2, MCM5, HIST1H1B, CDC6, MCM6, PCNA, BARD1, BRIP1, CHAF1A 
       DSN1, CDT1, FEN1, DHFR, CDC45, MCM2, HIST1H2AH, ZNF367, ATAD5, HIST1H4C 
Negative:  CCNB1, CDC20, AURKA, NEK2, KIF20A, CCNB2, PLK1, CDKN3, DLGAP5, PTTG1 
       DEPDC1, UBE2S, HMGB3, CENPE, PRR11, CENPA, BIRC5, HMMR, PIMREG, JPT1 
       KIF14, CDCA8, TROAP, PRC1, TPX2, KIF2C, KIF4A, CKS2, CDCA3, KNSTRN 

ElbowPlot(cocl2todabtram) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cocl2todabtram <- FindNeighbors(cocl2todabtram, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cocl2todabtram <- FindClusters(cocl2todabtram, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 1773
Number of edges: 55850

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8474
Number of communities: 10
Elapsed time: 0 seconds

cocl2todabtram <- RunUMAP(cocl2todabtram, dims = 1:15)

15:39:49 UMAP embedding parameters a = 0.9922 b = 1.112
15:39:49 Read 1773 rows and found 15 numeric columns
15:39:49 Using Annoy for neighbor search, n_neighbors = 30
15:39:49 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:39:50 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f83ea3e8f6
15:39:50 Searching Annoy index using 1 thread, search_k = 3000
15:39:50 Annoy recall = 100%
15:39:50 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:39:51 Initializing from normalized Laplacian + noise (using irlba)
15:39:51 Commencing optimization for 500 epochs, with 68614 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:39:54 Optimization finished

DimPlot(cocl2todabtram, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cocl2todabtram object
cocl2todabtram_input_data <- GetAssayData(cocl2todabtram, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cocl2todabtram_lin_pearson_list <- list()

for (i in fivecell_cDNA$CoCl2toDabTram){
  temp_pearson <- cor(cocl2todabtram_input_data[,colnames(cocl2todabtram)[cocl2todabtram$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cocl2todabtram_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cocl2todabtram_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cocl2todabtram_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CoCl2toDabTram){
    set.seed(j)
    num_cells <- length(cocl2todabtram$Lineage[cocl2todabtram$Lineage == i])
    temp_pearson <- cor(cocl2todabtram_input_data[,sample(colnames(cocl2todabtram), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cocl2todabtram_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cis

# Find the mean of the average pearson correlation per lineage
mean_pearson_cocl2todabtram <- mean(unlist(lapply(cocl2todabtram_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cocl2todabtram_sim <- sapply(1:length(cocl2todabtram_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cocl2todabtram_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cocl2todabtram <- (mean_pearson_cocl2todabtram-mean(means_pearson_cocl2todabtram_sim))/sd(means_pearson_cocl2todabtram_sim) # Z score comparing mean to simulations
pval_mean_pearson_cocl2todabtram <- pnorm(z_mean_pearson_cocl2todabtram, mean(means_pearson_cocl2todabtram_sim), sd(means_pearson_cocl2todabtram_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cocl2todabtram <- weighted.mean(unlist(lapply(cocl2todabtram_lin_pearson_list, mean)),
unlist(lapply(cocl2todabtram_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cocl2todabtram_sim <- sapply(1:length(cocl2todabtram_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cocl2todabtram_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cocl2todabtram_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cocl2todabtram <- (weighted_mean_pearson_cocl2todabtram-mean(weighted_means_pearson_cocl2todabtram_sim))/sd(weighted_means_pearson_cocl2todabtram_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cocl2todabtram <- pnorm(z_wmean_pearson_cocl2todabtram, mean(weighted_means_pearson_cocl2todabtram_sim), sd(weighted_means_pearson_cocl2todabtram_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cocl2todabtram <- c()
for (i in 1:length(cocl2todabtram_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cocl2todabtram_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cocl2todabtram <- cbind(wilcox_pval_cocl2todabtram, wilcox.test(x = unlist(lapply(cocl2todabtram_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cocl2todabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties

# Save outputs
save(cocl2todabtram, cocl2todabtram_lin_pearson_list, cocl2todabtram_lin_pearson_rand_list, z_mean_pearson_cocl2todabtram, pval_mean_pearson_cocl2todabtram, z_wmean_pearson_cocl2todabtram, pval_wmean_pearson_cocl2todabtram,  wilcox_pval_cocl2todabtram, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cocl2todabtram_pearson_sim_results.RData')
rm(cocl2todabtram, cocl2todabtram_lin_pearson_list, cocl2todabtram_lin_pearson_rand_list, cocl2todabtram_input_data)

Significance testing of the cis simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cis
cis <- subset(all_data, idents = 'cis') # Subset down to the cis object
cis <- NormalizeData(cis)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cis <- FindVariableFeatures(cis, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cis <- ScaleData(cis)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cis <- RunPCA(cis)

PC_ 1 
Positive:  GAS5, CRYAB, LINC00882, TYRP1, SPP1, LINC00520, SPACA3, ISG15, AL078612.2, IFI6 
       NUPR1, KRTAP19-1, DHRS2, LINC02269, TGFB2, IFIT2, CCL18, PRR9, LINC00492, IFIT1 
       AC006967.3, SERPINB2, AC021148.2, CIR1, AC103993.1, LINC00595, TRPM1, AL031429.1, AC109635.7, IGFL2-AS1 
Negative:  NEAT1, FUS, MKI67, PRKDC, TPI1, HNRNPAB, SFPQ, POLR2J3, FASN, ANKRD11 
       PRRC2C, HNRNPU, KHDRBS1, LINC00511, NUMA1, SSRP1, CCNI, MARCKSL1, HSP90B1, KPNB1 
       TIMP3, SLC25A13, MTDH, HNRNPA1, FADS1, CCT6A, HNRNPA2B1, SRSF4, GAS7, DNMT1 
PC_ 2 
Positive:  HSP90AA1, H2AFZ, PTMA, UBE2S, HMGB1, TPX2, MKI67, ACTG1, BIRC5, GAPDH 
       NUSAP1, ACTB, TK1, TUBB, LGALS1, TUBA1B, RANBP1, UBE2C, CKS1B, TOP2A 
       PTTG1, PRC1, RPS5, DBI, JPT1, NME1, TUBB4B, CFL1, ANP32E, HMGB2 
Negative:  MALAT1, GABPB1-AS1, NEAT1, CCDC144A, POLR2J3, AC008170.1, N4BP2L2, CADPS, SNHG14, MIR3142HG 
       ANKRD28, NKTR, TTN, KHDC4, AC058791.1, AUXG01000058.1, LINC-PINT, CBLB, RSRP1, CCNL1 
       HMGA2-AS1, PAXBP1, AC103691.1, STX16, FTX, SULT1C2, NHLRC3, WSB1, AC016831.5, LINC00488 
PC_ 3 
Positive:  ASPM, MKI67, CENPF, TOP2A, TPX2, GTSE1, KIF14, CENPE, UBE2C, AURKA 
       NUSAP1, DEPDC1, NUF2, ANLN, PRC1, KIF2C, CCNB1, PLK1, BUB1B, CDCA8 
       SGO2, CENPA, DLGAP5, HMGB2, KIF11, CEP55, CCNA2, HMMR, KNL1, HJURP 
Negative:  RPS8, GAPDH, RPS3A, RPS5, RPL5, RPS15A, RPL12, RACK1, RPL7A, SCD 
       BNIP3, CCNI, PGK1, ERBB3, RGS10, FGFBP2, EEF2, BTF3, NRN1, MARCKSL1 
       RPL23, IGFBP5, MIA, TPI1, MYC, APOD, PNRC1, HIST3H2A, BTG1, ADGRG1 
PC_ 4 
Positive:  HIST1H1B, HELLS, HIST1H1A, MCM4, DCT, BRCA1, GINS2, MCM10, CLSPN, FANCA 
       DTL, ATAD2, XRCC2, ATAD5, HIST1H1D, CHAF1A, C21orf58, HIST1H4C, MCM3, HIST1H1E 
       FRMD4B, BRCA2, EZH2, EXO1, CCNE2, MYBL2, FBXO5, LHFPL3-AS1, CCDC14, MCM2 
Negative:  FN1, SRGN, TMEM158, CAV1, ANXA1, STC1, ERRFI1, ARL4C, F2R, AHNAK 
       BASP1, SFRP1, IGFBP7, NGFR, VEGFA, LGALS1, JUN, CCND1, AC020916.1, FOSB 
       NDRG1, THBS1, AXL, LOXL2, ANXA2, MYOF, SERPINE2, TNFRSF12A, PRNP, SCG2 
PC_ 5 
Positive:  MCM4, CLSPN, ATAD2, UHRF1, GINS2, DTL, MCM3, HELLS, MCM6, E2F1 
       PCNA, HIST1H1B, MCM5, TMEM158, UNG, FN1, MCM10, CCNE2, CDCA7, CDC6 
       CHAF1A, MCM2, F2R, CDC25A, FANCA, POLD3, HIST1H1A, FOSB, RBBP8, MSH6 
Negative:  PLK1, CENPE, ASPM, CCNB1, HMMR, KIF14, GTSE1, ARL6IP1, AURKA, DLGAP5 
       CENPA, SGO2, TPX2, CCNB2, CDC20, NUSAP1, PRR11, CKS2, KIF20A, DEPDC1 
       NEK2, FAM83D, NUF2, CENPF, LHFPL3-AS1, CDCA8, KDM5B, TNS1, KIF2C, TOP2A 

ElbowPlot(cis) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cis <- FindNeighbors(cis, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cis <- FindClusters(cis, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 8303
Number of edges: 243168

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8429
Number of communities: 9
Elapsed time: 0 seconds

cis <- RunUMAP(cis, dims = 1:15)

15:40:48 UMAP embedding parameters a = 0.9922 b = 1.112
15:40:48 Read 8303 rows and found 15 numeric columns
15:40:48 Using Annoy for neighbor search, n_neighbors = 30
15:40:48 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:40:49 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f87b138e9a
15:40:49 Searching Annoy index using 1 thread, search_k = 3000
15:40:51 Annoy recall = 100%
15:40:51 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:40:52 Initializing from normalized Laplacian + noise (using irlba)
15:40:52 Commencing optimization for 500 epochs, with 370900 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:41:04 Optimization finished

DimPlot(cis, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cis object
cis_input_data <- GetAssayData(cis, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cis_lin_pearson_list <- list()

for (i in fivecell_cDNA$Cis){
  temp_pearson <- cor(cis_input_data[,colnames(cis)[cis$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cis_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cis_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cis_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$Cis){
    set.seed(j)
    num_cells <- length(cis$Lineage[cis$Lineage == i])
    temp_pearson <- cor(cis_input_data[,sample(colnames(cis), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cis_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cistocis

# Find the mean of the average pearson correlation per lineage
mean_pearson_cis <- mean(unlist(lapply(cis_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cis_sim <- sapply(1:length(cis_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cis_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cis <- (mean_pearson_cis-mean(means_pearson_cis_sim))/sd(means_pearson_cis_sim) # Z score comparing mean to simulations
pval_mean_pearson_cis <- pnorm(z_mean_pearson_cis, mean(means_pearson_cis_sim), sd(means_pearson_cis_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cis <- weighted.mean(unlist(lapply(cis_lin_pearson_list, mean)),
unlist(lapply(cis_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cis_sim <- sapply(1:length(cis_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cis_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cis_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cis <- (weighted_mean_pearson_cis-mean(weighted_means_pearson_cis_sim))/sd(weighted_means_pearson_cis_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cis <- pnorm(z_wmean_pearson_cis, mean(weighted_means_pearson_cis_sim), sd(weighted_means_pearson_cis_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cis <- c()
for (i in 1:length(cis_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cis_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cis <- cbind(wilcox_pval_cis, wilcox.test(x = unlist(lapply(cis_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

# Save outputs
save(cis, cis_lin_pearson_list, cis_lin_pearson_rand_list, z_mean_pearson_cis, pval_mean_pearson_cis, z_wmean_pearson_cis, pval_wmean_pearson_cis,  wilcox_pval_cis, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cis_pearson_sim_results.RData')
rm(cis, cis_lin_pearson_list, cis_lin_pearson_rand_list, cis_input_data)

Significance testing of the cistocis simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cistocis
cistocis <- subset(all_data, idents = 'cistocis') # Subset down to the cistocis object
cistocis <- NormalizeData(cistocis)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistocis <- FindVariableFeatures(cistocis, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistocis <- ScaleData(cistocis)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cistocis <- RunPCA(cistocis)

PC_ 1 
Positive:  MKI67, TPX2, CENPF, UBE2S, HMGB2, TUBB, TOP2A, TUBA1B, ASPM, ANLN 
       NUSAP1, PRC1, HMGB1, UBE2C, GTSE1, TMPO, BIRC5, TUBB4B, H2AFZ, AURKA 
       DTYMK, SMC4, AURKB, KPNA2, RRM2, KIF11, CENPE, CCNB1, HIST1H4C, CEP55 
Negative:  SAT1, GAS5, PNRC1, CBLB, SNHG14, GABPB1-AS1, HIST3H2A, MALAT1, PMEL, RPS8 
       BTG1, CREBRF, CCDC18-AS1, TRMT9B, A2M, NEAT1, LINC00492, APOE, SLC5A3, TNS1 
       SGCD, CD48, LINC01531, RACK1, PCDH7, RPL7A, LHFPL3-AS1, CADPS, NUPR1, CTSK 
PC_ 2 
Positive:  MLANA, PMEL, LGALS3, CHCHD6, LHFPL3-AS1, PRDX1, CD63, GSTO1, MIF, FRMD4B 
       AKR1A1, ATOX1, DCT, H2AFZ, SCD, CAPG, MITF, ADGRG1, TSPAN10, MIA 
       TIMM50, RLBP1, QPCT, CAPN3, RHOQ, BCAS3, FXYD3, METTL9, SIRPA, FABP5 
Negative:  FN1, IGFBP7, MYOF, THBS1, TMEM158, F2R, CAV1, BASP1, PRNP, SFRP1 
       ANXA2, SCG2, DKK1, AHNAK, ANXA1, SERPINE2, MMP2, LMO7, COL1A1, IL6ST 
       CALD1, CRIM1, ARL4C, VCL, DPYSL2, COL6A2, SORBS2, MMP1, AXL, ITGA2 
PC_ 3 
Positive:  MALAT1, NEAT1, TTN, PLAC4, ASPM, CCDC144A, LINC00488, AC058791.1, HMGA2-AS1, CLCN7 
       SYNE2, CENPF, AC008170.1, PRDM7, AUXG01000058.1, MIR3142HG, ZFYVE16, SNHG14, ANKRD11, DST 
       MKI67, RSRP1, NABP1, KIAA2026, NAV2, CBLB, AC016831.1, AC003681.1, KIF14, SMCR5 
Negative:  MIF, TMSB10, FTL, FTH1, UQCRH, DBI, SH3BGRL3, PSMA7, LGALS1, S100A6 
       ATP5MC3, RPS8, SERF2, RPL8, ATP5PF, NDUFC2, RPS6, TXN, MT2A, GYPC 
       S100A11, AP2S1, TMSB4X, COX7A2, NDUFB2, LGALS3, ATOX1, S100A13, ACTB, POMP 
PC_ 4 
Positive:  MCM4, MCM3, MCM6, HELLS, ATAD2, DTL, PCNA, BRCA1, CDC6, UHRF1 
       CLSPN, MCM5, GINS2, MSH6, UNG, CHAF1A, E2F1, CCNE2, POLD3, MCM7 
       MCM2, MCM10, HIST1H1B, FEN1, XRCC2, EXO1, CDCA7, WDR76, HIST1H1D, FANCA 
Negative:  CCNB1, AURKA, PLK1, HMMR, ARL6IP1, CENPE, CDC20, DLGAP5, KIF14, CENPA 
       PTTG1, TPX2, NEK2, GTSE1, ASPM, CCNB2, CDKN3, CKS2, PRR11, DEPDC1 
       KIF20A, UBE2C, UBE2S, KIF2C, BUB1, TUBB4B, NUF2, CEP55, BIRC5, CDCA8 
PC_ 5 
Positive:  TMEM158, SLC20A1, STC1, IGFN1, SCG2, MAGI1, SFRP1, NEAT1, VEGFA, RAB27B 
       COL6A5, NRP1, TFPI2, E2F7, MMP1, PRDM7, SRGN, ITGA3, IER3, SPOCK1 
       COL1A1, NPAS2, BASP1, ARSG, DCBLD2, XRCC2, BIRC7, LINC00488, LHFPL3-AS1, AC008170.1 
Negative:  PMP22, S100B, ANXA2, SPARC, SORBS2, PLP1, NIBAN1, PALLD, MARCKS, EPB41L3 
       AHNAK, TFAP2A, ALCAM, ADAM23, HSPG2, OLFML2A, NTRK2, IL17D, ESRP1, RPS8 
       SNAI2, CALD1, GAS7, PMP2, RHOBTB3, SLITRK6, MAP2, ZEB2, PYCARD, DYNLL1 

ElbowPlot(cistocis) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cistocis <- FindNeighbors(cistocis, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cistocis <- FindClusters(cistocis, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 3323
Number of edges: 112104

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8576
Number of communities: 11
Elapsed time: 0 seconds

cistocis <- RunUMAP(cistocis, dims = 1:15)

15:42:25 UMAP embedding parameters a = 0.9922 b = 1.112
15:42:25 Read 3323 rows and found 15 numeric columns
15:42:25 Using Annoy for neighbor search, n_neighbors = 30
15:42:25 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:42:25 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f840f75f96
15:42:25 Searching Annoy index using 1 thread, search_k = 3000
15:42:26 Annoy recall = 100%
15:42:26 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:42:27 Initializing from normalized Laplacian + noise (using irlba)
15:42:27 Commencing optimization for 500 epochs, with 131628 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:42:32 Optimization finished

DimPlot(cistocis, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cistocis object
cistocis_input_data <- GetAssayData(cistocis, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cistocis_lin_pearson_list <- list()

for (i in fivecell_cDNA$CistoCis){
  temp_pearson <- cor(cistocis_input_data[,colnames(cistocis)[cistocis$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cistocis_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cistocis_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cistocis_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CistoCis){
    set.seed(j)
    num_cells <- length(cistocis$Lineage[cistocis$Lineage == i])
    temp_pearson <- cor(cistocis_input_data[,sample(colnames(cistocis), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cistocis_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cistodabtram

# Find the mean of the average pearson correlation per lineage
mean_pearson_cistocis <- mean(unlist(lapply(cistocis_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cistocis_sim <- sapply(1:length(cistocis_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cistocis_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cistocis <- (mean_pearson_cistocis-mean(means_pearson_cistocis_sim))/sd(means_pearson_cistocis_sim) # Z score comparing mean to simulations
pval_mean_pearson_cistocis <- pnorm(z_mean_pearson_cistocis, mean(means_pearson_cistocis_sim), sd(means_pearson_cistocis_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cistocis <- weighted.mean(unlist(lapply(cistocis_lin_pearson_list, mean)),
unlist(lapply(cistocis_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cistocis_sim <- sapply(1:length(cistocis_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cistocis_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cistocis_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cistocis <- (weighted_mean_pearson_cistocis-mean(weighted_means_pearson_cistocis_sim))/sd(weighted_means_pearson_cistocis_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cistocis <- pnorm(z_wmean_pearson_cistocis, mean(weighted_means_pearson_cistocis_sim), sd(weighted_means_pearson_cistocis_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cistocis <- c()
for (i in 1:length(cistocis_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cistocis_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cistocis <- cbind(wilcox_pval_cistocis, wilcox.test(x = unlist(lapply(cistocis_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistocis_lin_pearson_list,  :
  cannot compute exact p-value with ties

# Save outputs
save(cistocis, cistocis_lin_pearson_list, cistocis_lin_pearson_rand_list, z_mean_pearson_cistocis, pval_mean_pearson_cistocis, z_wmean_pearson_cistocis, pval_wmean_pearson_cistocis,  wilcox_pval_cistocis, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cistocis_pearson_sim_results.RData')
rm(cistocis, cistocis_lin_pearson_list, cistocis_lin_pearson_rand_list, cistocis_input_data)

Significance testing of the cistodabtram simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cistodabtram
cistodabtram <- subset(all_data, idents = 'cistodabtram') # Subset down to the cistodabtram object
cistodabtram <- NormalizeData(cistodabtram)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistodabtram <- FindVariableFeatures(cistodabtram, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistodabtram <- ScaleData(cistodabtram)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cistodabtram <- RunPCA(cistodabtram)

PC_ 1 
Positive:  BASP1, CAV1, TUBB, MKI67, ANLN, PKM, CALR, DTYMK, TK1, UBE2S 
       TMPO, COL1A1, COTL1, SMC4, TPX2, TPM1, HIST1H4C, BIRC5, C12orf75, HSPD1 
       PCLAF, PRC1, NUSAP1, GTSE1, H2AFZ, TUBA1C, MYH9, CENPF, NXN, TUBA1B 
Negative:  PHACTR1, CRYAB, MLANA, SOX10, S100B, AKAP12, PMEL, ZNF704, PLP1, GPM6B 
       GAS7, SOX6, COL9A3, CITED1, CADPS, CBLB, ERBB3, SOX5, MXI1, AL139383.1 
       ID4, CA8, AC110285.1, TTN, TSC22D1, NFATC2, TFAP2A, IRS2, STARD4-AS1, AC068587.4 
PC_ 2 
Positive:  MKI67, CENPF, ANLN, TPX2, NUSAP1, GTSE1, ASPM, CEP55, PRC1, UBE2C 
       TOP2A, RRM2, STMN1, NCAPG, CENPE, S100B, BIRC5, KIF2C, CCNB1, CRYAB 
       HMMR, KNL1, PTTG1, KIF11, ITGA6, AURKB, MAD2L1, AURKA, CDCA2, GAS7 
Negative:  COL14A1, ITGB8, COL1A1, SLC12A8, COL12A1, IGFBP5, C1R, LINC00968, FTL, COL15A1 
       PDE5A, COL6A1, VCAM1, C1S, TXNRD1, IL6ST, CAMK2N1, SPON2, IGFBP7, DAB2 
       TMEM47, DEPTOR, SLC7A11, NUPR1, ITGA11, COL6A2, MSC, BGN, ADD3, NXN 
PC_ 3 
Positive:  TTN, AUXG01000058.1, CENPF, CCNB1, AC058791.1, HMGA2-AS1, BIRC5, CENPE, EREG, CEP55 
       CDC20, AURKA, AP000462.2, CDCA8, TPX2, DEPDC1, CENPA, AC114760.2, TOP2A, DLGAP5 
       MKI67, ANLN, PTTG1, PRC1, RRM2, UBE2C, ASPM, PCLAF, NEK2, CDKN3 
Negative:  HSPG2, EVI5, SPARC, IFI6, PALLD, OLFML2A, GAS7, COL9A3, PMP22, MFGE8 
       NFATC2, ANXA2, CTSD, GREM2, MMP2, AEBP1, ITGA2, DAG1, TUBA1A, CSRP2 
       FN1, MCAM, LIMCH1, SOX4, FAM89A, P4HA1, ARID5B, CD59, GPM6B, ITGA6 
PC_ 4 
Positive:  CPA4, ITGA3, FRMD4A, OXTR, DKK1, SCG5, SERPINE1, ARL4C, AC092807.3, KRT34 
       AXL, LMO7, SMYD3, SERINC2, CRISPLD2, TNIK, HMGA2, TPM1, LINC01638, PLPP4 
       TIMP3, GRAMD2B, VEGFC, MAGI1, PHLDB2, BDNF, PRSS23, TNC, SRGN, RPS27L 
Negative:  VCAM1, IGFBP5, GCLM, COL14A1, SLC7A11, FTL, C1R, CEBPD, TMEM47, ASPH 
       PGD, PITX1, EPS8, C1S, SRXN1, SOX4, PKDCC, TRIM16L, OSGIN1, HSD17B2 
       RHOBTB3, FOXF1, SQSTM1, DCN, ITGB8, LINC01914, AKR1B10, LINC00968, ITGA4, LSAMP 
PC_ 5 
Positive:  NEAT1, GLS, HMGA2-AS1, AUXG01000058.1, AC083870.1, CCDC14, AC058791.1, TTN, STARD4-AS1, AC012349.1 
       PZP, NABP1, AC114760.2, AP000462.2, CALD1, SORBS2, PSD3, HELLS, FANCA, IGFN1 
       XRCC2, BRCA1, C21orf58, AC016831.1, MAGI1, ATAD5, THBS1, COL1A1, FN1, BRCA2 
Negative:  PRDX1, LGALS3, CSTB, TMSB10, SH3BGRL3, FTL, S100A11, TMSB4X, MLANA, S100A10 
       CAPG, FKBP1A, MT2A, S100A6, CAV1, PTTG1, CCDC85B, NME1, CITED1, EZR 
       PKM, TRIM63, JPT1, AP1S2, APOE, TXN, HSPD1, H2AFZ, DSTN, SFRP1 

ElbowPlot(cistodabtram) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cistodabtram <- FindNeighbors(cistodabtram, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cistodabtram <- FindClusters(cistodabtram, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 2615
Number of edges: 89129

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8682
Number of communities: 9
Elapsed time: 0 seconds

cistodabtram <- RunUMAP(cistodabtram, dims = 1:15)

15:43:27 UMAP embedding parameters a = 0.9922 b = 1.112
15:43:27 Read 2615 rows and found 15 numeric columns
15:43:27 Using Annoy for neighbor search, n_neighbors = 30
15:43:27 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:43:27 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f830a4922c
15:43:27 Searching Annoy index using 1 thread, search_k = 3000
15:43:28 Annoy recall = 100%
15:43:28 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:43:29 Initializing from normalized Laplacian + noise (using irlba)
15:43:29 Commencing optimization for 500 epochs, with 105536 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:43:32 Optimization finished

DimPlot(cistodabtram, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cistodabtram object
cistodabtram_input_data <- GetAssayData(cistodabtram, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cistodabtram_lin_pearson_list <- list()

for (i in fivecell_cDNA$CistoDabTram){
  temp_pearson <- cor(cistodabtram_input_data[,colnames(cistodabtram)[cistodabtram$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cistodabtram_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cistodabtram_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cistodabtram_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CistoDabTram){
    set.seed(j)
    num_cells <- length(cistodabtram$Lineage[cistodabtram$Lineage == i])
    temp_pearson <- cor(cistodabtram_input_data[,sample(colnames(cistodabtram), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cistodabtram_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}

Look at whether lineages cluster together in each individual condition - cistococl2

# Find the mean of the average pearson correlation per lineage
mean_pearson_cistodabtram <- mean(unlist(lapply(cistodabtram_lin_pearson_list, mean))) # True mean of average correlations per lineage

means_pearson_cistodabtram_sim <- sapply(1:length(cistodabtram_lin_pearson_rand_list), function (y)
  mean(unlist(lapply(cistodabtram_lin_pearson_rand_list[[y]], mean)))) # list of mean of average correlations per lineage

z_mean_pearson_cistodabtram <- (mean_pearson_cistodabtram-mean(means_pearson_cistodabtram_sim))/sd(means_pearson_cistodabtram_sim) # Z score comparing mean to simulations
pval_mean_pearson_cistodabtram <- pnorm(z_mean_pearson_cistodabtram, mean(means_pearson_cistodabtram_sim), sd(means_pearson_cistodabtram_sim), lower.tail = F) # calculate p value from z score


# Find the weighted means of the average pearson correlations per lineage
weighted_mean_pearson_cistodabtram <- weighted.mean(unlist(lapply(cistodabtram_lin_pearson_list, mean)),
unlist(lapply(cistodabtram_lin_pearson_list, length))) # true weighted mean of average correlations per lineage

weighted_means_pearson_cistodabtram_sim <- sapply(1:length(cistodabtram_lin_pearson_rand_list), function(y)
  weighted.mean(unlist(lapply(cistodabtram_lin_pearson_rand_list[[y]], mean)),
                unlist(lapply(cistodabtram_lin_pearson_rand_list[[y]], length)))) # List of weighted means of pearson correlations

z_wmean_pearson_cistodabtram <- (weighted_mean_pearson_cistodabtram-mean(weighted_means_pearson_cistodabtram_sim))/sd(weighted_means_pearson_cistodabtram_sim) # Z score comparing mean to simulations
pval_wmean_pearson_cistodabtram <- pnorm(z_wmean_pearson_cistodabtram, mean(weighted_means_pearson_cistodabtram_sim), sd(weighted_means_pearson_cistodabtram_sim), lower.tail = F) # calculate p value from z score

# Compare each individual distribution of pearson correlations to the observed correlation by wilcoxon rank sum test and track pval
wilcox_pval_cistodabtram <- c()
for (i in 1:length(cistodabtram_lin_pearson_rand_list)){
  sim_means <- unlist(lapply(cistodabtram_lin_pearson_rand_list[[i]], mean))
  wilcox_pval_cistodabtram <- cbind(wilcox_pval_cistodabtram, wilcox.test(x = unlist(lapply(cistodabtram_lin_pearson_list, mean)),
                                                                y = sim_means, alternative = 'greater')$p.value)
}

Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties
Warning in wilcox.test.default(x = unlist(lapply(cistodabtram_lin_pearson_list,  :
  cannot compute exact p-value with ties

# Save outputs
save(cistodabtram, cistodabtram_lin_pearson_list, cistodabtram_lin_pearson_rand_list, z_mean_pearson_cistodabtram, pval_mean_pearson_cistodabtram, z_wmean_pearson_cistodabtram, pval_wmean_pearson_cistodabtram,  wilcox_pval_cistodabtram, file = '2022_01_14_analysis_scripts/2022_05_27_analysis/Lineage_expression/cistodabtram_pearson_sim_results.RData')
rm(cistodabtram, cistodabtram_lin_pearson_list, cistodabtram_lin_pearson_rand_list, cistodabtram_input_data)

Significance testing of the cistococl2 simulation

Idents(all_data) <- all_data$OG_condition # Change the idents to the OG condition for subsetting to cistococl2
cistococl2 <- subset(all_data, idents = 'cistococl2') # Subset down to the cistococl2 object
cistococl2 <- NormalizeData(cistococl2)

Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistococl2 <- FindVariableFeatures(cistococl2, selection.method = 'vst', nFeatures = 20000)

Warning: The following arguments are not used: nFeatures
Calculating gene variances
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
Calculating feature variances of standardized and clipped values
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

cistococl2 <- ScaleData(cistococl2)

Centering and scaling data matrix

  |                                                                                                 
  |                                                                                           |   0%
  |                                                                                                 
  |==============================================                                             |  50%
  |                                                                                                 
  |===========================================================================================| 100%

cistococl2 <- RunPCA(cistococl2)

PC_ 1 
Positive:  MALAT1, HMGA2-AS1, TTN, CSTB, AC058791.1, LINC00488, PLAC4, SNHG14, MIR3142HG, AC008170.1 
       NEAT1, AC008771.1, IGFL2-AS1, HSPA1B, TM4SF19-AS1, PRR26, NABP1, HIF1A-AS3, SAT1, HSPA1A 
       LINC01705, AL356599.1, AL121772.3, AL162426.1, MIR22HG, AL022323.4, ARSG, UCKL1-AS1, CLEC2D, KCNC1 
Negative:  CFL1, PRKDC, PKM, RPS8, ACTB, CAV1, SFRP1, TIMP3, KDELR1, TPI1 
       CALR, HSP90B1, YWHAZ, HIST1H4C, MKI67, HNRNPD, HNRNPAB, MTDH, ITGB1, AP2S1 
       CCNI, EEF2, HIST1H1B, RCN1, NCL, PA2G4, CCT6A, ACTN1, LDHB, MCM4 
PC_ 2 
Positive:  PMEL, MLANA, MT-ND4, H2AFZ, MT-CYB, MITF, MYH10, DCT, MT-ND1, ATOX1 
       CHCHD6, S100B, GPM6B, CAPG, EEF1A1, MT-ND2, SMS, MKI67, LDHB, MT-ND3 
       NBL1, PIK3R3, RPS8, RPS15A, HMGB1, MIA, RPL28, FRMD4B, PARP1, FABP5 
Negative:  MT2A, MALAT1, TMEM158, SFRP1, FN1, SCG2, IER3, TFPI2, IL6ST, EREG 
       VEGFA, BASP1, JUN, ITGA2, PHLDA1, PPP1R15A, ARL4C, IGFBP7, LUCAT1, DKK1 
       SRGN, SIRPB1, TNFRSF12A, SERPINB2, ITGA3, STC1, GNG11, COL1A1, HIF1A-AS3, PLAUR 
PC_ 3 
Positive:  FTL, FTH1, TMSB10, LGALS1, RPL28, S100A6, MT2A, TMSB4X, RPS15A, RPS8 
       EEF1A1, SH3BGRL3, PRDX1, RACK1, NQO1, NDUFA4, CSTB, ACTG1, ATOX1, CFL1 
       RAB13, COX8A, UQCRH, ACTB, GSTP1, H2AFZ, CCND1, GAS5, EEF1D, RPL12 
Negative:  NEAT1, MALAT1, ANKRD11, ZFYVE16, GOLGA4, GABPB1-AS1, HMGA2-AS1, FRMD4B, SNHG14, MIR3142HG 
       AC058791.1, NFKBIZ, LINC02249, PLAC4, VMP1, ARSG, AC016831.1, AC068587.4, CADPS, SLC5A3 
       SLC20A1, CBLB, AKAP12, NABP1, AC008170.1, PIK3R3, FRMD4A, TRIM25, AC008771.1, AF117829.1 
PC_ 4 
Positive:  SFRP1, COL1A1, CXCL12, COL6A2, COL6A1, WNT5A, IGFBP7, SRGN, MT-ND4, CCN4 
       SCG2, MT-CYB, FN1, EREG, MT-ND1, LTBP1, TMEM158, NRG1, MT-ND3, NRP1 
       S100A6, TIMP3, MT-ND2, IGFBP5, AQP1, SIRPB1, SPOCK1, LYPD6, ITGA11, SNAP25 
Negative:  GADD45B, PPP1R15A, DDIT3, IER2, SNHG7, H3F3B, HSPA1A, EIF5, HSPH1, KBTBD8 
       HSPA1B, SERTAD1, ZFAND2A, BRD2, ATF4, DNAJB9, AL118516.1, ODC1, ATF3, AC003092.1 
       PDRG1, LRIF1, SNHG12, BUD31, LINC00520, SNHG15, DNAJB1, OSER1, ATP6V0B, IL24 
PC_ 5 
Positive:  RPS8, SERPINE2, PMP22, LY6E, PLP1, SPARC, RAMP1, GPM6B, MIA, VKORC1 
       SORBS2, CCNI, CEBPD, CST3, SEZ6L2, MFSD12, CADM1, PRSS35, KDELR1, TFAP2A 
       AHNAK, RPL28, MARCKS, GAS7, SLC5A3, CDH19, CBLB, CADPS, SLC44A1, MARCKSL1 
Negative:  MKI67, CENPF, ASPM, TOP2A, UBE2C, NUSAP1, ANLN, GTSE1, TPX2, HMGB2 
       RRM2, CEP55, AURKA, PRC1, AURKB, CENPE, KIF14, HIST1H1B, SMC4, NCAPG 
       SGO2, KIF23, HJURP, KNL1, CCNB1, ATAD2, DEPDC1, KIF11, CDK1, BIRC5 

ElbowPlot(cistococl2) # The standard deviation seems to really level off at 10

# Recluster with the appropriate number of dimensions
cistococl2 <- FindNeighbors(cistococl2, dims = 1:15)

Computing nearest neighbor graph
Computing SNN

cistococl2 <- FindClusters(cistococl2, resolution = 0.5)

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 5715
Number of edges: 198390

Running Louvain algorithm...

0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

Maximum modularity in 10 random starts: 0.8451
Number of communities: 6
Elapsed time: 0 seconds

cistococl2 <- RunUMAP(cistococl2, dims = 1:15)

15:46:18 UMAP embedding parameters a = 0.9922 b = 1.112
15:46:18 Read 5715 rows and found 15 numeric columns
15:46:18 Using Annoy for neighbor search, n_neighbors = 30
15:46:18 Building Annoy index with metric = cosine, n_trees = 50
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:46:19 Writing NN index file to temp file /var/folders/ph/24prrxys02179y9_qzhxjgvc0000gn/T//Rtmpx16qOy/filec5f84789fa7
15:46:19 Searching Annoy index using 1 thread, search_k = 3000
15:46:20 Annoy recall = 100%
15:46:20 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
15:46:21 Initializing from normalized Laplacian + noise (using irlba)
15:46:21 Commencing optimization for 500 epochs, with 238026 positive edges
Using method 'umap'
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|
15:46:29 Optimization finished

DimPlot(cistococl2, reduction = 'umap', pt.size = 1)

# Get the scaled data from the cistococl2 object
cistococl2_input_data <- GetAssayData(cistococl2, assay = 'RNA', slot = 'scale.data')

# Build list of lineages with at least 5 cells in dab tram and do pearson correlation
cistococl2_lin_pearson_list <- list()

for (i in fivecell_cDNA$CistoCoCl2){
  temp_pearson <- cor(cistococl2_input_data[,colnames(cistococl2)[cistococl2$Lineage == i]])
  temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
  cistococl2_lin_pearson_list[[i]] <- temp_pearson_filt
}

# Need to do a random sampling of the same thing 
cistococl2_lin_pearson_rand_list <- list()
num_iter <- 100 
for(j in 1:num_iter){
  cistococl2_lin_pearson_rand_list[[j]] <- list()
  for (i in fivecell_cDNA$CistoCoCl2){
    set.seed(j)
    num_cells <- length(cistococl2$Lineage[cistococl2$Lineage == i])
    temp_pearson <- cor(cistococl2_input_data[,sample(colnames(cistococl2), num_cells, replace = F)])
    temp_pearson_filt <- temp_pearson[lower.tri(temp_pearson, diag = FALSE)]
    cistococl2_lin_pearson_rand_list[[j]][[i]] <- temp_pearson_filt
    
  }
}